#Recommendation Systems

[callmeadzy]

Types of Recommendation Systems :

Collaborative Filtering : It's like when your buddies exchange their favourite restaurant secrets, so you can avoid eating at places that serve questionable concoctions that Gordon Ramsey would say ‘ A Mouth Full of Hubba-Bubba ‘! Content-Based Filtering : Imagine a personal fashion guru who looks at your outfits and suggests clothes that match your style, saving you from fashion disasters that could make Lady Gaga cringe!! Hybrid Recommendation : It's like combining two smoothie flavors - collaborative filtering and content-based filtering - to create a blend that caters to your taste buds, resulting in a recommendation feast fit for a hungry hippo.. Knowledge-Based Filtering : Picture a wise owl (Hedwig’s brother) who knows all your desires and interests, guiding you through a vast library of books, movies, and hobbies, just like a friendly librarian who doesn't shush you for laughing too loud.. Context-Aware Filtering : It's like having a trusty sidekick who tailors recommendations based on your surroundings, making sure you don't accidentally watch a romantic comedy when you're in the middle of the Hannibal movie marathon.

[scatterbrain-lion]

i think i get how recommendation systems, or literally how any similarity retrieval system works. the general pipeline always goes like:

normalize numerical attributes, cleanup categorical attributes and text descriptions (gather more numerical and categorical attributes from descriptions like NER)

vectorize categorical and numerical attributes

create text and image embeddings (pretrained resnet, sbert, bidirectional lstm, the methods vary depending on what you're getting)

concatenate the categorical-numerical-text-image vectors into 1 content vector space

you could either knn right there on the content vector space and you'll probably get a decent similarity retrieval system if you plan on taking like top 5 or 10

if you wanna make sure you're getting the absolutely best items for your retrieval and you have ground truth data for what you wanna retrieve and not retrieve, then try a siamese neural network (contrastic loss or triple loss)

get cosine similarity from new embeddings in the content vector space, determine positive and negative similarity to find an appropriate cutoff point for retrieval

maybe i should go into recommendation systems instead of computer vision, that might actually align closer to studying towards mlops or ml infrastructure

[thewordharbor]

How recommendation systems work using AI

Ever wondered why Netflix and Spotify feel like they know your taste?

Did you know that recommendation systems are behind those tailored suggestions you see on platforms like Netflix, Amazon, and Spotify? They’re the reason your favorite shows, products, or playlists feel like they’re handpicked just for you.  But how exactly do these systems work? The magic happens through artificial intelligence, where algorithms analyze your behavior to predict what you’ll like…

[realjdobypr]

Unlock AI-Powered Topic Recommendations for Targeted Traffic

The Role of Data in AI-Powered Recommendations Harnessing the Power of Data for Personalized Suggestions In the era of digital transformation, data serves as the cornerstone for driving AI-powered recommendations. Through the analysis of user behavior, preferences, and historical data, businesses can derive invaluable insights to offer personalized suggestions. This not only enriches the user…

[atissi]

kras mazov lookin ass

[ahmedhelmy-stuff-blog]

Technology: Amazing Transformations and Astounding Impacts on Daily Life

Information and communication technology has become an integral part of our daily lives and is evolving at a rapid and continuous pace. This topic includes many fields such as artificial intelligence, robotics, new applications, smart devices, and other developments in technology. Artificial intelligence is one of the most common topics in technology, where AI-based systems analyze data and…

[qwikskills]

Neural networks are a powerful tool for machine learning and artificial intelligence. They are inspired by the structure and function of the human brain and have revolutionized the way computers can learn from data and make predictions.

At their core, neural networks consist of interconnected nodes, or artificial neurons, that work together to analyze data and make predictions. They are capable of learning from vast amounts of data and can be trained to identify patterns and make predictions with a high degree of accuracy.

Neural networks have a wide range of applications, including image and speech recognition, natural language processing, and recommendation systems. They can be used in a variety of industries, including healthcare, finance, and retail, to make data-driven decisions and improve processes.

If you're interested in exploring the world of neural networks, there are many resources available, including online courses and tutorials, books, and forums. With the right resources and a bit of patience, anyone can learn how to use neural networks to make predictions and solve complex problems.

In conclusion, neural networks are a powerful tool for machine learning and artificial intelligence. Whether you're just starting out or looking to expand your skills, learning about neural networks is a valuable investment that can open up new opportunities and help you succeed in the rapidly evolving world of technology.

[felikatze]

i feel like. the more i like and care about something. the less i am capable of watching video essays about it. game i never heard of and don't intend to ever play? sure i'll watch 8hrs discussing it's flaws.

but thing i like? if you think you can point out flaws i'm not already aware of, you are dead wrong. none know better how much my interests suck than me.

and also. if you get one thing wrong about them i'll maul you. with things i like it means i've already seen every single piss on the poor take of it ever, and i'm much more polarized. i got emotional investment. i'm going to start biting people.

[piosplayhouse]

Cue Olivia Newton-John Physical

Equality version:

[non-dys-sys]

It’s never okay to lock a headmate away, never. They could be the worst person in the world, you still shouldn’t lock them away. There’s a difference between keeping headmates away from eachother and taking away someone’s privacy, free will and connections. You are abandoning that headmate, you are telling them that they are not worth helping, that they don’t deserve to get better. If talking it out doesn’t work for them there are other options, I’d recommend giving them their own space to speak their mind without threat of mistreatment for it, for starters. If someone really doesn’t want help don’t force it on them, just let them know it’s always an option and leave it at that. -Ange

[irrideemaple]

"Now that I'm here, there's no more point to you"

[callmeadzy]

RecSys

A recommendation system is like a psychic pizza delivery guy who magically predicts your craving for pineapple and anchovy toppings. Meme on!

[maalidoesart]

ray of sunlight on us

[illuminchim]

Shout out to the person who sent me an ask, wondering if I could sketch Shen Qingqiu. I answered it in private and now I can't find it 🫠

I had a lot of fun drawing him and currently, I'm finding the time to color it. To the person that asked I hope you see this!

[wubbelwubbwubb]

Seriously, I don’t know why I didn’t just say you’re welcome and please get out of my cubicle so I can sit here and leak in peace.

[muffinlance]

hellooo very cool person I am just now getting into the atla fandom and I know Nothing other than your fics are very cool and great and I was wondering if you have any recs (or links to posts with recs!!) bc I trust your judgment LOL

I hope this is ok to ask!! also I will add that I am not that picky but I will add that I am very much a longfic enjoyer so🫶🫶

My friend, may I open up to you the broad world of clicking an AO3 user's bookmarks. <3

AKA: literally click any ao3 username, "bookmarks" should appear towards the top of the resulting page. You can then voyage into the additionally wonderful waters of filtering by length, "recommended", fandom, etc.

Also: if the fic you like is in collections, try checking them out, especially if the title appeals to you. Can be a great way to find essentially a fic playlist.

Anyway all hail ao3's designers they done good work