#Context Clusters in Search Query Suggestions

[anthonyjohn01]

How to find and use keywords that trigger AI Overview

In today’s new reality of search, ranking on the top of page one of Google is just not good enough anymore. AI-generated summaries don’t simply choose the most popular answer. They’re not triggered by any exact keyword phrases but rather kind of broader keyword strings and content type.

What Determines When a Keyword Triggers an AI Introduction

In SEO, not all keywords are created equal. Many are just too broad or too general to make it to enough of a specific response to land a spot in Google’s new AI response box. The most common keywords that make AI Overviews trigger are generally the ones associated with questions that are informational, instructional or very niche and specific.

Google’s AI has learned to interpret these kinds of statements as invites for assistance and that’s when AI can jump in, providing a synthesized response.

Platforms like Seobix are at the forefront of providing real time intent analysis with keyword suggestions that go beyond the basics, helping creators navigate these more nuanced changes in user behaviour.

Getting to the Bottom of User Intent for AI-Friendly Queries

AI Overviews are made possible by Google’s mission to deliver the highest quality answer to a query ; rather than the most highly optimized web page. That’s because Google’s algorithm is filtering for pages most directly answering the intent of the query.

By analysing what users are actually searching for and zeroing in on how these searchers type or speak in natural language, you can discover the types of keyword phrases that will set you apart from the pack and directly or indirectly deliver what Google wants. Tools like Seobix, break down these patterns through the analysis of SERP features in real-time and AI answer trends, allowing you to focus your efforts on related keyword clusters of high-opportunity content that are more likely to evoke AI summaries.

Producing Material Worth the AI

Finding the right keyword is just part of the battle. To truly show up in the AI Overview, your material has to be laid out in a clear format that AI can digest and extract information from. That translates to descriptive, concise headlines, putting the most relevant answer early on in your content, and surrounding it with context that helps prove your expertise on the subject.

The balance of your content can then add layers of depth to that central explanation, offering examples, context, other tools.

Tracking AI Opportunities in Real Time

Google Search Console has just started rolling out new metrics to let users see impressions and clicks from AI Overviews but it’s very limited data so far.

This is where a tool like Seobix comes in and adds real value. By surfacing keyword opportunities, tracking AI visibility trends, and providing AI-optimized content recommendations, it makes it easy to keep your SEO strategy in-line with the way search is going to work

Conclusion

Google’s all-knowing AI doesn’t operate on keywords ; it operates on meaning. Therefore, as content creators, it is our responsibility to go beyond niche little catch phrases and start constructing based on theme-based relevance. The keywords that activate AI Overviews are much less about quantity and more about specificity, purpose, and value. By focusing on the best search intent-focused phrases and developing content that aligns with them, you improve your ability to rank and you improve your chances to be featured.

[avisheksen-1993]

7 Smart Ways Digital Marketers Are Using Google LLMs to Win in 2025

7 Smart Ways Digital Marketers Are Using Google LLMs to Win in 2025

Introduction: 

I’m Avishek Sen, a professional digital marketer, and let me tell you something honestly—2025 is not about working harder, it’s about working smarter with AI. And one of the biggest game-changers right now? Google’s LLMs (Large Language Models), especially Gemini.

Here’s a fact that says it all:

🧠 Over 80% of digital marketers in 2025 are now using AI tools to plan, create, and optimize content — and Google LLMs are leading the pack. (Source: HubSpot AI Trends Report 2025)

These AI tools can understand natural language, generate blog content, write ad copy, suggest SEO keywords, analyze user intent, and even give voice-based insights — all within seconds.

Still think AI is only for tech experts?

📊 57% of beginner marketers (including those just 12th pass or graduate-level students) are already using tools like Google Gemini to improve their marketing game. (Source: Think with Google, 2025)

And guess what? It’s working.

📈 Businesses using AI-powered content creation and SEO optimization have seen a 40% increase in engagement and up to 2x more conversions compared to those using traditional methods.

So if you're still doing everything manually—content, SEO, research, analytics—you’re spending too much time and missing out on better results.

In this blog, I’ll show you 7 smart and proven ways digital marketers are winning in 2025 with the help of Google’s LLMs. I’ve used these same strategies with my students, freelancers, and small business clients—and the results have been amazing.

1. Supercharged Content Creation in Minutes

Writing high-quality blogs, ad copy, and social media posts used to take hours. Now? With Google Gemini, marketers can generate content outlines, full articles, and captions in seconds.

🔹 Stat: Content creation time has dropped by 65% with the help of LLMs in 2025. (Source: Content Marketing Institute)

Why it works:

Google LLMs understand context, tone, and audience intent.

You can create multilingual content easily.

Gemini helps you maintain brand voice across platforms.

Example: A student in my 6-month course created a full 1500-word SEO blog with Gemini in 10 minutes and ranked it on Google within a week.

2. Smarter Keyword Research and SEO Strategy

Keyword research is no longer just about volume; it's about intent and relevance.

Google LLMs can analyze your niche, scan search trends, and give you keyword clusters that match user behavior patterns.

🔹 Stat: 73% of SEO professionals in 2025 use AI tools like Google Gemini for keyword research. (Source: SEMrush 2025 Survey)

What it helps with:

Long-tail keyword discovery

Featured snippet targeting

Competitor keyword gap analysis

Bonus Tip: Combine Gemini's output with tools like Google Search Console or Ahrefs for powerful SEO wins.

3. Personalized Ad Copy That Converts

Ad fatigue is real in 2025. That’s why LLMs are now used to create dynamic ad copy tailored to audience segments.

💼 Facebook & Google Ads using AI-generated copy have seen conversion rates improve by 32%. (Source: AdEspresso AI Benchmark Report)

Benefits:

A/B test headlines and descriptions

Customize copy for retargeting

Maintain emotional tone and urgency

Real Story: One of my clients used LLM-generated ad variants for a lead-gen campaign and reduced their CPC by 40% while increasing click-through rate (CTR).

4. Advanced Customer Insights with Natural Language Queries

You no longer need to be a data scientist to read analytics.

Just ask Google LLMs simple questions like:

"Which blog got the highest engagement last month?"

"What type of audience is converting most in Delhi?"

📊 Businesses using LLMs for customer insights are 32% more efficient in campaign targeting. (Source: MarketingProfs AI Insights)

LLMs can summarize analytics data and even suggest actions. This means faster decision-making and more accurate targeting.

5. Building Powerful Email Campaigns with AI

Emails are still gold. But writing subject lines, body copy, and call-to-actions (CTAs) that convert? That takes skill.

Now with Google LLMs:

Generate catchy subject lines

Write personalized email sequences

Segment emails based on customer behavior

📧 AI-personalized emails are seeing a 50%+ open rate compared to traditional bulk campaigns. (Source: Mailchimp Trends 2025)

Tip: Use Gemini to write 3 variations of the same email for A/B testing. It’s fast, free, and effective.

6. Faster Social Media Management and Engagement

Social media needs daily content, trends, and engagement. LLMs can help you:

Schedule engaging posts

Respond with AI-drafted comments and replies

Follow trends and hashtags

📊 64% of social media marketers say AI tools have improved their productivity by at least 2 hours per day. (Source: Sprout Social 2025)

Use Cases:

Instagram caption generator

Twitter thread planning

Comment reply assistant

Client Example: A bakery in Kolkata grew their Instagram page from 300 to 3,200 followers in 60 days using Gemini-written captions and smart posting strategies.

7. Creating Better Video Scripts and Voiceovers

Video is king. And Google LLMs are helping you rule it.

You can now:

Write short reels scripts in seconds

Translate video content into multiple languages

Create emotional storytelling using data-driven input

🎥 LLM-generated video content ideas lead to 23% higher watch time on platforms like YouTube Shorts and Instagram Reels. (Source: HubSpot Video Trends 2025)

Bonus: You can even use Gemini to generate subtitles, call-to-actions, and thumbnail text ideas.

Why Smart Marketers Are Adopting AI – And You Should Too

Whether you're a student, freelancer, or small business owner, AI isn't here to steal your job. It’s here to help you grow faster.

❌ Don't ignore AI tools because you feel they're "too technical." 

✅ Embrace tools like Google LLMs to automate the boring stuff and focus on your creativity.

Digital marketing in 2025 is about balance:

AI + Human = Super Marketer

Speed + Strategy = Success

As someone who trains students daily, I’ve seen beginners go from zero to freelance income just by mastering these tools.

Conclusion: Embrace the Future of Digital Marketing Today

The future of digital marketing is here, and it’s powered by smart AI tools like Google’s LLMs. In 2025, success is no longer about how hard you work but how smart you work. From creating content in minutes to crafting personalized ads, unlocking deep customer insights, and managing social media with ease — these AI tools are revolutionizing the way marketers operate.

If you want to stay competitive, save time, and boost your results, embracing Google LLMs isn’t optional — it’s essential. The combination of AI’s speed and your creativity can unlock new growth opportunities and help you build a thriving digital presence faster than ever before.

Whether you’re a beginner, a freelancer, or a business owner, these tools will give you the edge to win in today’s fast-moving digital world. And if you’re in Kolkata and looking for the best digital marketer to guide you through this AI-powered transformation, I’m here to help.

Get in touch with me, Avishek Sen — your trusted digital marketing expert in Kolkata. Let’s create winning strategies together and make 2025 your most successful year yet.

WhatsApp me anytime: +91 90071 66621 or you can mail me avishek@seosrv.com

Written by Avishek Sen Professional Digital Marketer | SEOSrv

[anthony2231]

How Google Trends Can Instantly Upgrade Your SEO Game

Keyword research is no longer about stuffing content with popular search terms. In 2025, it's about understanding what users truly want and building content that meets their search intention. The good news is: you do not need a big budget or complicated software to make it right.

With the right free SEO tools, a little strategy, followed by a consistent workflow, an SEO foundation can be laid down to channel traffic and grow an online presence—whether it is a blog, business site, or eCommerce store.

The Importance of Keywords in 2025

Keywords have remained the foremost user-content linking point despite AI innovations and algorithm changes. The big difference is Google knowing context and intent more than ever now.

So rather than simply saying best running shoes, it is important to find the answers to:

Who is searching?

What are they trying to achieve?

What kind of content are they expecting?

That is when keyword research in SEO becomes the crux. It is not so much about words-it is about the right words.

Step-By-Step Guide for Smarter Keyword Research

Start With Real Questions

Use a free keyword research tool to find out what your audience is asking. Long-tail keywords such as 'best running shoes for flat feet' or 'budget-friendly trail running shoes' usually generate more qualified traffic than broad competitive terms.

 Analyze the Data Behind the Words

Use a SEO analyzer to assess:

Keyword difficulty

Search volume

SERP competition

User intent

This step ensures you're not wasting time targeting impossible-to-rank terms.

Run a Free SEO Audit

Before optimizing new pages, run a free SEO audit on your existing content. This will help uncover on-page issues, like missing headers, poor keyword distribution, or technical flaws that could drag your rankings down.

 Build Keyword Clusters

Don’t stop at one keyword. Group related terms together. For example:

“best free SEO tool for bloggers”

“SEO tools for beginners”

“best SEO optimization tool 2025”

This allows you to write comprehensive, topic-rich content that ranks for multiple queries at once.

Top Tools That Make It Easy (And Free)

Here are the must-have tools that’ll make your research process fast, efficient, and budget-friendly:

Tool Type Function

Free SEO Tools Discover keywords, analyze volume, and trends

SEO Analyzer Examine site health, keyword placement, and structure

Free SEO Audit Find technical issues holding your site back

Best SEO Optimization Tool Refine content structure, tags, and internal linking

Together, these tools help you build a content strategy that actually performs.

Why This Method Is Perfect for Startups & Solo Creators

Whether you're a startup with limited resources or a solo content creator, this strategy gives you access to powerful SEO without paying for expensive software. The best free SEO tools provide nearly everything you need:

Real-time keyword suggestions

On-page SEO tips

Competitor research

Link audits

No fluff. Just real results.

Avoid These Keyword Mistakes in 2025

Even with the best tools, it’s easy to fall into common traps:

Targeting keywords that are too broad

Ignoring search intent

Skipping technical optimization

Overusing keywords unnaturally

Use a free SEO audit and SEO analyzer to catch these issues before they hurt your performance.

Final Thoughts

SEO in 2025 rewards those who are strategic, not just prolific. The most successful websites are those that combine smart keyword planning, strong technical SEO, and relevant content.

So, if you’re ready to grow your organic traffic, don’t guess. Use the best SEO tool, run a free SEO audit, and apply keyword analysis in SEO that’s built around real user intent.

[digitalswarna]

Digital Marketing Training Institute in Hyderabad 

The SEO landscape is shifting faster than ever before, and at the center of this transformation is Artificial Intelligence (AI). Gone are the days of manually sifting through keyword lists, blindly targeting search volume, or guessing what your audience wants. In 2025, AI is completely redefining keyword research—making it smarter, faster, and more accurate.

Whether you’re a digital marketer, SEO professional, content creator, or small business owner, understanding how AI is shaping keyword research is no longer optional—it’s essential. In this article, we’ll explore how AI tools, algorithms, and smart assistants are revolutionizing the way we discover and use keywords to rank and convert.

✅ The Old Way vs. The AI Way

Traditional Keyword Research:

Manual input into tools like Google Keyword Planner

Focus on search volume, CPC, and competition

Long lists of isolated keywords

Guessing user intent

Static data snapshots

AI-Driven Keyword Research:

Predictive keyword suggestions based on trends and user behavior

Smart keyword clustering and topic modeling

Real-time, intent-based suggestions

NLP (Natural Language Processing) for understanding context

Dynamic updates and insights powered by machine learning

AI doesn't just give you keywords—it gives you strategies.

✅ 1. Smarter Keyword Suggestions

In the past, keyword tools simply returned a list of related terms based on what you typed in. AI tools today go several steps further by:

Understanding context around your topic

Using machine learning to predict what users might search for next

Analyzing massive datasets in real-time to uncover rising search terms

Suggesting long-tail keywords and questions based on how users talk and search (especially important for voice SEO)

Tools Leading the Way:

Frase and Surfer SEO analyze top-ranking pages and recommend keyword-rich content outlines

ChatGPT-based custom GPTs can brainstorm topic clusters, FAQs, and keyword ideas in seconds

Ubersuggest AI predicts keyword trends over time

✅ 2. Better Understanding of Search Intent

Google's algorithms now reward content that aligns with search intent, not just keywords. That’s where AI truly shines. It can analyze:

User behavior across search engines

The type of content that currently ranks (blogs, videos, product pages)

Semantic patterns in language

By recognizing if a keyword is informational, navigational, transactional, or investigative, AI tools help you create the right type of content to match what users actually want.

Example:

Searching “best email marketing tools”? AI identifies it as a commercial investigation keyword, suggesting a comparison-style blog post.

✅ 3. Keyword Clustering and Topic Modeling

Instead of targeting isolated keywords, SEO pros in 2025 focus on content clusters and topical authority. AI makes this incredibly efficient through:

Automatic keyword grouping: Tools like Surfer SEO, Scalenut, and MarketMuse group related terms based on semantic similarity

Topic modeling: AI uncovers related subtopics and questions to cover in your content

Internal linking suggestions: AI even tells you how to connect content pieces to strengthen SEO

This not only improves rankings—it builds authority in your niche.

✅ 4. Voice Search and Conversational Queries

With over half of all searches now being voice-based, AI plays a huge role in adapting keyword research for this format. Unlike text searches, voice searches are:

Longer

More conversational

Question-based

AI adapts by:

Using NLP to detect spoken query patterns

Suggesting natural-sounding keywords and question formats

Helping you create FAQ-rich content that targets featured snippets and “People Also Ask” sections

Example:

Text: “best pizza NYC” Voice: “Where can I get the best pizza in New York City right now?”

AI ensures you’re optimizing for both.

✅ 5. Real-Time Trend Detection

Traditional keyword tools provide monthly or quarterly search data. But AI tools can detect real-time trends by analyzing:

Social media chatter

News and blog trends

Search behavior shifts

Tools like:

Exploding Topics and Google Trends with AI enhancements

Custom AI dashboards that integrate real-time search data

AI assistants that scan Reddit, Twitter, and YouTube for hot topics

This enables marketers to jump on trends before they go mainstream—giving you the first-mover advantage in content creation.

✅ 6. Personalized Keyword Research Based on Niche and Audience

AI isn’t just fast—it’s personal. In 2025, keyword research tools can now:

Learn your industry, audience, and brand tone

Recommend keywords that fit your specific customer journey

Filter out irrelevant keywords and surface high-converting ones

Some AI tools even connect with your Google Analytics and Search Console data to tailor recommendations based on what’s actually working on your site.

✅ 7. Predictive Analytics and SEO Forecasting

Want to know which keywords will matter next quarter or next year? AI-powered SEO platforms now offer predictive keyword analytics, helping you plan ahead by forecasting:

Search trends

Seasonal keyword performance

Emerging queries in your niche

This is a game-changer for editorial calendars, product launches, and long-term content strategies.

✅ 8. Automating the Research Workflow

Before AI, keyword research was a time-consuming manual task. Now, AI tools automate the entire workflow:

Input a seed topic → Get keyword clusters, intent analysis, suggested content types, and outline recommendations within minutes.

Integrate with your CMS (like WordPress or Webflow) to push content plans directly into your blog backend

Use AI chat assistants to brainstorm, refine, and validate your keyword strategies on the fly

Automation = faster output and better accuracy.

✅ What This Means for SEO Professionals

If you’re in SEO or content marketing, embracing AI in your keyword research means:

✅ Faster results ✅ Smarter targeting ✅ Better content alignment ✅ Stronger ROI

But it also means the bar is higher. Generic keyword lists and outdated tactics just won’t cut it anymore. Your competitors are likely using AI—and if you’re not, you’re already behind.

Final Thoughts: The Future of Keyword Research Is Here

AI isn’t a threat to SEO professionals—it’s a powerful partner. In 2025, those who embrace AI for keyword research will:

Create better content

Rank higher

Adapt faster to trends

Save time and scale smartly

So if you haven’t explored AI-driven keyword research yet, now’s the time. Because in the SEO game of 2025, smart is fast—and AI is smarter than ever.

[techsavvy-agcy]

Decoding the Google Core Update, March 2025

Decoding the Google Core Update, March 2025

On March 13th, 2025, the constantly changing and unforeseeable digital world felt another major shakeup with the release of Google's newest core algorithm update. Just like before, the official explanation of the update was typically unclear, highlighting the continued focus on "rewarding top-notch content" and "enhancing search accuracy." Still, the impact was evident, causing a lot of worry and hurried changes in the SEO field.

This piece explores the effects, some educated guesses, and the best ways to handle the Google Core update from March 13th, 2025. It aims to clarify how the algorithm is changing and what that means for those who run websites and create content.

The Aftermath: Search Results in Flux

Right after the update, the search engine results pages (SERPs) saw some major shifts. Tools that monitor such volatility showed huge jumps, pointing to big changes in rankings across different sectors. Initial reports zeroed in on a few key areas:

E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness): Google has always cared a lot about E-E-A-T in how they judge website quality, but the update in March 2025 seemed to turn up the volume on it. Websites that showed they knew their stuff, had proven expertise, were authoritative, and were super trustworthy did a lot better.

Content Authenticity and Originality: Google seemed to get much stricter about content that AI just cranked out without much original thought or real value. Sites that used a bunch of rehashed content, rewritten articles, or stuff that looked like it came straight from an AI template saw their rankings take a nosedive.

User Experience (UX) and Page Experience: Things like how fast your site loads, if it's easy to use on mobile, and overall how user-friendly it is were way more important after the update. Slow websites, had annoying pop-ups or were a pain to use on phones got dinged.

Contextual Relevance and Semantic Understanding: Google's update appeared to fine-tune its capacity for grasping the context and purpose embedded within search queries. Sites that provided comprehensive, in-depth content that addressed the nuances of user searches saw improvements.

Data Driven Content: Websites that used first party data to personalize and improve user experiences, and that used data to back up claims, saw significant improvements.

Analyzing the Algorithm's Evolving Priorities

The Rise of "Human-Centric" Content: The update reinforced the notion that Google is prioritizing content created by humans, for humans. This goes beyond simply avoiding AI-generated text; it emphasizes the importance of genuine insights, personal experiences, and a conversational tone.

Emphasis on Real-World Experience: The "Experience" element of E-E-A-T took center stage. This suggests that Google is placing greater value on content creators who have real-world experience in the topics they cover. First hand accounts and demonstrations of real world product usage were rewarded.

Combatting Misinformation and Disinformation: In an era of rampant misinformation, Google appeared to be strengthening its defenses against unreliable or misleading content. This involved stricter scrutiny of sources, a greater emphasis on fact-checking, and a preference for content from established authorities.

The Importance of Contextual Understanding: The upgrade demonstrated Google's developing capacity to understand linguistic nuances and the context of user enquiries. This shows that semantic SEO and topic clustering are becoming more important.

Privacy and User Data: With growing user concern about privacy, Google's algorithm appears to be rewarding sites with clear data-gathering policies that prioritize customer privacy.

Video and Multimedia: Video content, especially content that is original, well produced, and engaging, got a large boost in rankings. Long form, well edited videos, that are not just re-purposed content, were very well received by the algorithm.

Emerging Best Practices for the Post-March 2025 Landscape

In the wake of the update, website owners and content creators were forced to adapt their strategies. The following best practices emerged as essential for navigating the evolving search landscape:

Prioritize E-E-A-T: Focus on increasing your brand's authority and trustworthiness. Display your knowledge through author profiles, credentials, and real-world experience. Seek expert contributions and endorsements.

Create Original, High-Quality Content: Invest in original research, in-depth analysis, and distinctive opinions. Avoid using AI-generated content that lacks originality or value.

Optimize the user experience:  Make sure your website is speedy, mobile-friendly, and simple to navigate. Address Core Web Vitals concerns and prioritize a consistent user experience.

Embrace semantic SEO: Create content that takes into account the complexities of user intent and search query context. Topic clustering and semantic keywords can help you develop topical authority.

Build Authentic Backlinks: Concentrate on obtaining high-quality backlinks from credible sources. Avoid using black-hat link-building strategies.

Focus on Data and Transparency:  Show users how data is gathered and used. Ensure that privacy policies are explicit and user data is secure.

Invest in Video Content: Make high-quality, engaging videos that add value to your viewers. Optimize videos for search and make them accessible.

Monitor and Adapt: Stay up to current on algorithm updates, and regularly monitor the performance of your website. Prepare to adapt your strategies as needed.

Focus on First-Party Data: Collect and use first-party data to tailor the customer experience. This enables more targeted content and a better experience.

Create a Community:  Create a user community centered around your brand. This can be accomplished using social media, forums, and other online channels. Strong communities generate social signals, which Google values.

The long-term implications

The Google core upgrade on March 13th, 2025, marked a substantial shift in the algorithm's objectives, emphasizing the significance of human-centric information, genuine expertise, and a user-first approach. As Google improves its comprehension of language and user intent, website owners and content providers must adapt to the changing standards.

The update is likely to have the following long-term implications:

A More Level Playing Field for Smaller Websites: By emphasizing E-E-A-T and original material, Google may level the playing field for smaller websites that can demonstrate actual expertise and offer distinctive value.

A Decline in Low-Quality, AI-Generated Content: The crackdown on AI-generated content is expected to result in a decrease in low-quality, spammy content that clutters the SERPs.

A Greater Emphasis on User Experience: Google's continuous emphasis on UX is likely to result in overall improvements in website design and performance.

A More Contextually Relevant Search Experience: Google's improved semantic understanding is expected to result in a more relevant and fulfilling search experience for users.

A renewed focus on the human element of content creation: The value of real people, and their real-world experiences, will only continue to increase.

In conclusion, the March 13th, 2025, Google core update served as a stark reminder of the dynamic nature of the digital landscape. By prioritizing E-E-A-T, originality, user experience, and contextual understanding, Google is pushing website owners and content creators to embrace a more human-centric and value-driven approach. Those who can adapt to these evolving standards will be well-positioned to thrive in the years to come. The future of search is focused on genuine human connection and genuine human experience.

[seoexpert264]

SEO Experts Discuss How to Optimize for Google's BERT Algorithm

1. Focus on Natural Language and User Intent

Google's BERT (Bidirectional Encoder Representations from Transformers) algorithm is designed to understand natural language and the context of search queries more effectively. SEO experts recommend optimizing content by focusing on user intent and creating natural, conversational language that answers questions in a way that aligns with how people actually search. Rather than stuffing content with keywords, aim to address the underlying question or need of the user.

Related Post: SEO expert in Bangladesh

2. Create Long-Form, In-Depth Content

BERT excels at understanding long-tail, conversational queries, and it helps rank content that provides detailed, in-depth answers. SEO professionals suggest creating high-quality, comprehensive content that fully explores topics and answers specific questions. The more detailed and thorough your content, the more likely it is to match user queries and rank higher in search results.

3. Improve Content Clarity and Structure

Since BERT focuses on understanding context and meaning, content clarity and structure are critical for SEO success. SEO experts recommend using clear and concise language, as well as breaking down complex ideas into easy-to-understand sections. Use headings, bullet points, and short paragraphs to improve readability and ensure that your content is structured in a way that search engines can easily interpret.

4. Optimize for Featured Snippets

BERT’s advanced understanding of natural language has made featured snippets more important than ever. SEO experts advise optimizing for featured snippets by directly answering common questions in your content. Use concise, well-formatted answers in the form of lists, tables, or Q&A sections to increase the chances of your content being selected as a featured snippet.

Related Post: SEO expert in Bangladesh

5. Use Semantic SEO Techniques

BERT focuses on understanding the meaning behind words, so SEO experts recommend using semantic SEO techniques. This involves creating content that is contextually relevant and incorporates related terms, synonyms, and topic clusters. By covering all aspects of a topic and using semantically related keywords, you can increase the likelihood of your content ranking for a wider range of search queries.

6. Optimize for Mobile and Voice Search

Since BERT improves the understanding of conversational queries, it aligns well with voice search, which is typically more natural and question-based. SEO professionals suggest optimizing your site for mobile and voice search by ensuring it loads quickly, has a mobile-responsive design, and contains conversational phrases or questions that people might ask through voice assistants like Google Assistant.

7. Regularly Update Content

BERT continues to evolve, and SEO experts recommend keeping your content up-to-date with the latest information. Regularly updating your blog posts and website content ensures that you stay relevant and maintain high-quality answers to user queries, which can help improve your rankings under BERT’s algorithm.

By implementing these strategies, SEO experts recommend that businesses optimize their content to align with Google's BERT algorithm, improving rankings, relevance, and visibility in search results.

Related Post: SEO expert in Bangladesh

[eternalelevator]

Mastering Semantic Search: How to Optimize for User Intent

It's no longer about stuffing keywords into content; it's about understanding and aligning with user intent. In this blog, we’ll delve into advanced strategies for optimizing your content and website for semantic search, providing actionable tips to keep you ahead in 2025.

What is Semantic Search?

Semantic search is the process by which search engines understand the context and intent behind a query, rather than focusing solely on matching keywords. Powered by AI and machine learning, semantic search delivers more accurate, context-aware results, considering factors like:

Searcher context: Location, device, past behavior.

Entity recognition: Identifying people, places, and concepts.

Natural language understanding: Interpreting synonyms, queries, and questions.

This evolution has shifted the focus from individual keywords to clusters of related concepts.

Why Optimizing for User Intent Matters

User intent—the goal behind a search query—is at the heart of semantic search. By aligning your content with the four key types of user intent, you can:

Informational: Provide answers to “How”, “What”, and “Why” questions.

Navigational: Help users find specific websites or brands.

Transactional: Support purchase decisions.

Commercial investigation: Offer in-depth comparisons and insights.

The better your content meets these intents, the higher your chances of ranking.

Advanced Strategies for Semantic Search Optimization

1. Implement Topic Clustering

Move beyond standalone keywords and structure your content around topic clusters. This involves creating:

Pillar pages: Comprehensive content pieces covering core topics.

Cluster content: Supporting articles that delve into specific subtopics and link back to the pillar page.

This approach signals to search engines that your site is an authority on a subject, improving relevance and rankings.

2. Leverage Schema Markup

Schema markup helps search engines understand your content's context. Use structured data to enhance rich snippets, such as:

Product reviews and ratings.

FAQs and how-to guides.

Event details.

Adding schema boosts visibility and click-through rates (CTR) by making your results stand out.

3. Focus on Natural Language Processing (NLP)

Craft content that mirrors how people naturally search and speak. To do this:

Use conversational tones and question-answer formats.

Target long-tail keywords and questions.

Analyze People Also Ask (PAA) boxes for trending queries.

4. Utilize AI-Powered Tools

AI-driven tools like Google’s Natural Language API or MarketMuse can:

Analyze your content’s topical relevance.

Identify gaps in your keyword and entity coverage.

Suggest semantic improvements.

5. Optimize for Voice Search

With voice search on the rise, focus on:

Answering common voice queries succinctly.

Optimizing for local search terms (e.g., “near me” queries).

Using structured data to ensure compatibility with voice assistants.

6. Enhance E-A-T Signals

Google emphasizes expertise, authoritativeness, and trustworthiness (E-A-T) in rankings. To strengthen E-A-T:

Showcase credentials for authors.

Earn backlinks from reputable sites.

Regularly update content to maintain accuracy and relevance.

7. Invest in Visual Search Optimization

With advancements in image recognition, optimizing for visual search can give you an edge:

Use descriptive alt texts and filenames.

Add structured data for images.

Optimize image load speeds.

8. Track Search Intent Changes

User intent evolves. Regularly review your analytics to identify shifts in search behavior. Use tools like SEMrush and Ahrefs to monitor trending queries and adapt your content strategy accordingly.

Measuring Success

Evaluate the effectiveness of your semantic search optimization by:

Monitoring CTR and bounce rates.

Tracking keyword rankings for topic clusters.

Measuring engagement metrics like time on page and social shares.

Conclusion

Mastering semantic search in 2025 requires a forward-thinking approach. By focusing on user intent, leveraging AI tools, and optimizing for natural language and visual search, you can stay ahead in the ever-evolving SEO landscape. Embrace these strategies, and you’ll not only improve rankings but also create meaningful connections with your audience.

[dostudio]

How do I find thousands of keywords to rank No. 1 on Google?

Ranking No. 1 on Google for thousands of keywords is a challenging task that requires a comprehensive and strategic approach to keyword research, content creation, and optimization. Here's a step-by-step guide to help you find and target a large number of keywords:

1. Define Your Niche and Topics:

Identify your niche or industry and the main topics relevant to your business. Having a clear understanding of your target audience and offerings is crucial.

2. Seed Keywords:

Create a list of seed keywords related to your niche. These are broad terms that represent the main topics you want to cover.

3. Use Keyword Research Tools:

Leverage keyword research tools to expand your list. Tools such as Google Keyword Planner, Ubersuggest, SEMrush, Ahrefs, and others can provide additional keyword suggestions, search volumes, and competition levels.

4. Long-Tail Keywords:

Explore long-tail keywords that are more specific and detailed. Long-tail keywords often have lower competition and can attract more targeted traffic. Look for variations, question-based queries, and specific user intents.

5. Competitor Analysis:

Analyze competitors in your industry to identify keywords they are targeting. This can provide insights into gaps in your keyword strategy or opportunities for differentiation.

6. Content Cluster Strategy:

Implement a content cluster strategy. Group related keywords into clusters, and create pillar content pages that cover the main topic comprehensively. Support these pillar pages with cluster pages that target specific long-tail keywords.

7. Semantic Keyword Exploration:

Consider semantic keywords and related terms that are conceptually linked to your main topics. Search engines use semantic analysis to understand the context and relevance of content.

8. Use Google Autocomplete and Related Searches:

Utilize Google Autocomplete by entering your seed keywords and examining the suggested queries. Additionally, scroll to the bottom of Google search results to find related searches.

9. Location-Specific Keywords:

If your business has a local presence, include location-specific keywords to target users in specific geographic areas.

10. User Intent Analysis:

sqlCopy code- Understand the intent behind each keyword. Are users looking for information, products, reviews, or local services? Align your content with user intent to improve relevance. 

11. Create Comprehensive Content:

cssCopy code- Develop comprehensive, high-quality content that addresses the needs and questions of your target audience. Each piece of content should be well-optimized for the keywords it targets. 

12. Regularly Update and Expand:

cssCopy code- Continuously update and expand your content. Regularly revisit your keyword strategy to identify new opportunities and adapt to changes in user behavior or industry trends. 

13. Optimize Technical SEO:

cssCopy code- Ensure that your website has sound technical SEO elements in place, such as proper site structure, clean URLs, and mobile optimization. Technical SEO plays a crucial role in overall search visibility. 

14. Backlink Strategy:

cssCopy code- Develop a backlink strategy to build authority for your website. High-quality backlinks from reputable sources can positively impact your rankings. 

15. Monitor and Adjust:

sqlCopy code- Regularly monitor the performance of your keywords using analytics tools. Adjust your strategy based on changes in search algorithms, user behavior, and your business goals. 

Remember that achieving No. 1 rankings for thousands of keywords requires time, effort, and ongoing optimization. Prioritize quality over quantity, and focus on providing valuable content that meets the needs of your audience. Regularly analyze your performance and make data-driven adjustments to your strategy.DO Studio : Best Marketing Agency In CalicutDo Studio is the Best Marketing agency in Calicut. We do Digital Marketing, Branding, Web Design, Web Developement, Package Design & Print Designhttps://dostudio.co.in/

[miettawilliemk1]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[wendyjudithqe]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[mariaajameso]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[rodrigueztha]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[lindasharonbn]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[janiceclaudetteo]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[miettawilliemk]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P

[mariathaterh]

Context Clusters in Search Query Suggestions

unsplash-logoSaketh Garuda

Context Clusters and Query Suggestions at Google

A new patent application from Google tells us about how the search engine may use context to find query suggestions before a searcher has completed typing in a full query. After seeing this patent, I’ve been thinking about previous patents I’ve seen from Google that have similarities.

It’s not the first time I’ve written about a Google Patent involving query suggestions. I’ve written about a couple of other patents that were very informative, in the past:

6/10/2016 – Google Entity Search Suggestions Patent (Associating an entity with a search query)

5/26/2010How a Search Engine Might Identify Possible Query Suggestions (Generating query suggestions using contextual information)

In both of those, the inclusion of entities in a query impacted the suggestions that were returned. This patent takes a slightly different approach, by also looking at context.

Context Clusters in Query Suggestions

We’ve been seeing the word Context spring up in Google patents recently. Context terms from knowledge bases appearing on pages that focus on the same query term with different meanings, and we have also seen pages that are about specific people using a disambiguation approach. While these were recent, I did blog about a paper in 2007, which talks about query context with an author from Yahoo. The paper was Using Query Contexts in Information Retrieval. The abstract from the paper provides a good glimpse into what it covers:

User query is an element that specifies an information need, but it is not the only one. Studies in literature have found many contextual factors that strongly influence the interpretation of a query. Recent studies have tried to consider the user’s interests by creating a user profile. However, a single profile for a user may not be sufficient for a variety of queries of the user. In this study, we propose to use query-specific contexts instead of user-centric ones, including context around query and context within query. The former specifies the environment of a query such as the domain of interest, while the latter refers to context words within the query, which is particularly useful for the selection of relevant term relations. In this paper, both types of context are integrated in an IR model based on language modeling. Our experiments on several TREC collections show that each of the context factors brings significant improvements in retrieval effectiveness.

The Google patent doesn’t take a user-based approach ether, but does look at some user contexts and interests. It sounds like searchers might be offered a chance to select a context cluster before showing query suggestions:

In some implementations, a set of queries (e.g., movie times, movie trailers) related to a particular topic (e.g., movies) may be grouped into context clusters. Given a context of a user device for a user, one or more context clusters may be presented to the user when the user is initiating a search operation, but prior to the user inputting one or more characters of the search query. For example, based on a user’s context (e.g., location, date and time, indicated user preferences and interests), when a user event occurs indicating the user is initiating a process of providing a search query (e.g., opening a web page associated with a search engine), one or more context clusters (e.g., “movies”) may be presented to the user for selection input prior to the user entering any query input. The user may select one of the context clusters that are presented and then a list of queries grouped into the context cluster may be presented as options for a query input selection.

I often look up the inventors of patents to get a sense of what else they may have written, and worked upon. I looked up Jakob D. Uszkoreit in LinkedIn, and his profile doesn’t surprise me. He tells us there of his experience at Google:

Previously I started and led a research team in Google Machine Intelligence, working on large-scale deep learning for natural language understanding, with applications in the Google Assistant and other products.

This passage reminded me of the search results being shown to me by the Google Assistant, which are based upon interests that I have shared with Google over time, and that Google allows me to update from time to time. If the inventor of this patent worked on Google Assistant, that doesn’t surprise me. I haven’t been offered context clusters yet (and wouldn’t know what those might look like if Google did offer them. I suspect if Google does start offering them, I will realize that I have found them at the time they are offered to me.)

Like many patents do, this one tells us what is “innovative” about it. It looks at:

…query data indicating query inputs received from user devices of a plurality of users, the query data also indicating an input context that describes, for each query input, an input context of the query input that is different from content described by the query input; grouping, by the data processing apparatus, the query inputs into context clusters based, in part, on the input context for each of the query inputs and the content described by each query input; determining, by the data processing apparatus, for each of the context clusters, a context cluster probability based on respective probabilities of entry of the query inputs that belong to the context cluster, the context cluster probability being indicative of a probability that at least one query input that belongs to the context cluster and provided for an input context of the context cluster will be selected by the user; and storing, in a data storage system accessible by the data processing apparatus, data describing the context clusters and the context cluster probabilities.

It also tells us that it will calculate probabilities that certain context clusters might be requested by a searcher. So how does Google know what to suggest as context clusters?

Each context cluster includes a group of one or more queries, the grouping being based on the input context (e.g., location, date and time, indicated user preferences and interests) for each of the query inputs, when the query input was provided, and the content described by each query input. One or more context clusters may be presented to the user for input selection based on a context cluster probability, which is based on the context of the user device and respective probabilities of entry of the query inputs that belong to the context cluster. The context cluster probability is indicative of a probability that at least one query input that belongs to the context cluster will be selected by the user. Upon selection of one of the context clusters that is presented to the user, a list of queries grouped into the context cluster may be presented as options for a query input selection. This advantageously results in individual query suggestions for query inputs that belong to the context cluster but that alone would not otherwise be provided due to their respectively low individual selection probabilities. Accordingly, users’ informational needs are more likely to be satisfied.

The Patent in this patent application is:

(US20190050450) Query Composition System Publication Number: 20190050450 Publication Date: February 14, 2019 Applicants: Google LLC Inventors: Jakob D. Uszkoreit Abstract:

Methods, systems, and apparatus for generating data describing context clusters and context cluster probabilities, wherein each context cluster includes query inputs based on the input context for each of the query inputs and the content described by each query input, and each context cluster probability indicates a probability that at a query input that belongs to the context cluster will be selected by the user, receiving, from a user device, an indication of a user event that includes data indicating a context of the user device, selecting as a selected context cluster, based on the context cluster probabilities for each of the context clusters and the context of the user device, a context cluster for selection input by the user device, and providing, to the user device, data that causes the user device to display a context cluster selection input that indicates the selected context cluster for user selection.

What are Context Clusters as Query Suggestions?

The patent tells us that context clusters might be triggered when someone is starting a query on a web browser. I tried it out, starting a search for “movies” and got a number of suggestions that were combinations of queries, or what seem to be context clusters:

The patent says that context clusters would appear before someone began typing, based upon topics and user information such as location. So, if I were at a shopping mall that had a movie theatre, I might see Search suggestions for movies like the ones shown here:

One of those clusters involved “Movies about Business”, which I selected, and it showed me a carousel, and buttons with subcategories to also choose from. This seems to be a context cluster:

This seems to be a pretty new idea, and may be something that Google would announce as an availble option when it becomes available, if it does become available, much like they did with the Google Assistant. I usually check through the news from my Google Assistant at least once a day. If it starts offering search suggestions based upon things like my location, it could potentially be very interesting.

User Query Histories

The patent tells us that context clusters selected to be shown to a searcher might be based upon previous queries from a searcher, and provides the following example:

Further, a user query history may be provided by the user device (or stored in the log data) that includes queries and contexts previously provided by the user, and this information may also factor into the probability that a user may provide a particular query or a query within a particular context cluster. For example, if the user that initiates the user event provides a query for “movie show times” many Friday afternoons between 4 PM-6 PM, then when the user initiates the user event on a Friday afternoon in the future between these times, the probability associated with the user inputting “movie show times” may be boosted for that user. Consequentially, based on this example, the corresponding context cluster probability of the context cluster to which the query belongs may likewise be boosted with respect to that user.

It’s not easy to tell whether the examples I provided about movies above are related to this patent or if it is tied more closely to the search results that appear in Google Assistant results. It’s worth reading through and thinking about potential experimental searches to see if they might influence the results that you may see. It is interesting that Google may attempt to anticipate what is suggests to show to us as query suggestions, after showing us search results based upon what it believes are our interests based upon searches that we have performed or interests that we have identified for Google Assistant.

The contex cluster may be related to the location and time that someone accesses the search engine. The patent provides an example of what might be seen by the searcher like this:

In the current example, the user may be in the location of MegaPlex, which includes a department store, restaurants, and a movie theater. Additionally, the user context may indicate that the user event was initiated on a Friday evening at 6 PM. Upon the user initiating the user event, the search system and/or context cluster system may access the content cluster data 214 to determine whether one or more context clusters is to be provided to the user device as an input selection based at least in part on the context of the user. Based on the context of the user, the context cluster system and/or search system may determine, for each query in each context cluster, a probability that the user will provide that query and aggregate the probability for the context cluster to obtain a context cluster probability.

In the current example, there may be four queries grouped into the “Movies” cluster, four queries grouped into the “Restaurants” cluster, and three queries grouped into the “Dept. Store” cluster. Based on the analysis of the content cluster data, the context cluster system may determine that the aggregate probability of the queries in each of the “Movies” cluster, “Restaurant” cluster, and “Dept. Store” cluster have a high enough likelihood (e.g., meet a threshold probability) to be input by the user, based on the user context, that the context clusters are to be presented to the user for selection input in the search engine web site.

I could see running such a search at a shopping mall, to learn more about the location I was at, and what I could find there, from dining places to movies being shown. That sounds like it could be the start of an interesting adventure.

Copyright © 2019 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately. Plugin by Taragana

The post Context Clusters in Search Query Suggestions appeared first on SEO by the Sea ⚓.

http://bit.ly/2Ne7Y6P