#Hybrid Neuromorphic Systems

[mehmetyildizmelbourne-blog]

Brainoware: The Hybrid Neuromorphic System for a Brighter Tomorrow

A glimpse into the double-edged nature of Brain Organoid Reservoir Computing, with the pros/cons of this biological computing approach From a young age, I was captivated by the mysteries of science and the promise of technology, wondering how they could shape our understanding of the world. I was fortunate to receive STEM education early on in a specialized school, where my creativity and…

[nestronautomation]

The Future of Computing: A New Era of Intelligence

The evolution of computing technology is accelerating at an unprecedented pace. Gone are the days when silicon-based CPUs ruled alone—today, we’re on the cusp of a diverse, integrated computing ecosystem combining quantum, neuromorphic, photonic, and biological systems.

1. Quantum Computing Goes Commercial

In the UK, firms like Oxford Quantum Circuits and Riverlane have introduced error-corrected quantum systems into commercial data centers—marking a foundational leap toward practical quantum use Times. Meanwhile, D‑Wave’s rising stock reflects growing business adoption, with over half of surveyed companies planning to integrate quantum tech within the next few years . Cloud providers are also making quantum power accessible via hybrid quantum‑classical models

2. Neuromorphic & Photonic Computing

Neuromorphic chips—hardware designed to mimic brain-like neural networks—are gaining traction. These systems operate at lower power and can adapt dynamically . Alongside this, photonic computing, using light instead of electricity, promises breakthroughs in AI and quantum data centers due to its superior heat and power efficiency

3. Ambient Intelligence & Edge Revolution

Computing is moving into our surroundings. Ambient intelligence embeds smart, invisible systems into homes, offices, and cities, responding adaptively to us. Combined with edge computing and rising 5G/6G networks, real-time data processing becomes practical on mobile devices, enabling autonomous vehicles, remote healthcare, and immersive AR/VR experiences

4. Security for a Quantum Age

As quantum power grows, current encryption becomes vulnerable. The field is responding with post-quantum cryptography—new algorithms designed to resist quantum attacks. Concurrently, zero-trust models and AI-powered defenses are becoming central to cloud-native infrastructure

5. Sustainable & Green Computing

Environmental concerns are shaping tech: energy-efficient chips, carbon-neutral data centers, and hardware longevity are now priorities. Cloud and AI providers are investing heavily in sustainable architectures to balance innovation with ecological responsibility

6. Biocomputing & Brain-Computer Hybrids

Emerging research into organoid intelligence explores using lab-grown brain tissue for computation—combining biology with electronics for ultra-efficient processing . In parallel, bioconvergence blends biotechnology and computation—leading to breakthroughs in brain‑machine interfaces, biochips, and regenerative systems .

🌐 What This Means for You

Unmatched compute power: Quantum and neuromorphic systems will transform drug discovery, climate modeling, and logistics.

Invisible yet intelligent environments: Expect your devices to become contextual companions, predicting needs instantly.

Stronger, greener security: Prepare for tech that safeguards data in a post-quantum world—secure, sustainable, smart.

Closer human-tech integration: Biocomputing and brain-inspired systems may revolutionize health, learning, and productivity.

🔭 Conclusion

The future of computing isn’t a singular evolution—it’s a mosaic. Quantum, neuromorphic, photonic, and biological systems will coexist, each excelling at different tasks. Their integration, powered by advanced networks and secure infrastructures, will redefine everything from medicine to smart cities. As this transformation gains momentum, organizations and individuals must prepare now—by adopting quantum-safe systems, investing in green tech, and embracing flexible architectures. We're not just entering a new chapter of computing; we're composing a multi-dimensional symphony of innovation.

to know more about visit garce it solutions

[komalllsinhh]

Photonic Neuron Chip Market - Key Players, Size, Trends, Growth Opportunities, Analysis and Forecast 

Photonic Neuron Chip Market, Trends, Business Strategies 2025-2032

The global Photonic Neuron Chip Market size was valued at US$ 78.34 million in 2024 and is projected to reach US$ 289.67 million by 2032, at a CAGR of 17.89% during the forecast period 2025–2032. This explosive growth trajectory reflects the increasing adoption of neuromorphic computing solutions across multiple industries.

Photonic neuron chips represent a breakthrough in artificial intelligence hardware, mimicking biological neural networks using light-based processing. These chips leverage photonic integrated circuits to perform high-speed, low-energy computations for applications including signal processing, data analysis, and image recognition. Unlike traditional electronic chips, photonic neuron chips offer significant advantages in terms of speed, bandwidth, and energy efficiency—critical factors driving market expansion.

The market growth is fueled by several key factors: rising demand for AI-powered solutions across sectors, increasing investments in quantum computing research, and the need for energy-efficient computing architectures. While the semiconductor industry faced challenges in 2022 with only 4.4% overall growth due to economic factors, photonic computing technologies continue gaining traction. Major players like Intel, IBM, and Samsung are actively developing photonic neuron solutions, with North America currently leading in adoption due to strong R&D infrastructure and government support for advanced computing initiatives.

Our comprehensive Market report is ready with the latest trends, growth opportunities, and strategic analysis. https://semiconductorinsight.com/download-sample-report/?product_id=103240

Segment Analysis:

By Type

Signal Processing Segment Dominates Due to High-Speed Computing Capabilities

The market is segmented based on type into:

Signal Processing

Subtypes: Optical signal conversion, noise reduction modules

Data Processing

Image Identification

Subtypes: Pattern recognition, object detection

Others

By Application

Telecommunications Segment Leads Owing to 5G Network Expansion

The market is segmented based on application into:

Aviation

Telecommunications

Automotive

Others

By Technology

Optical Neural Networks Gain Traction for Energy Efficiency Advantages

The market is segmented based on technology into:

Silicon Photonics

Optical Neural Networks

Hybrid Integration

By End User

Enterprise Segment Shows Strong Adoption for AI Applications

The market is segmented based on end user into:

Defense and Aerospace

Enterprise

Research Institutions

Others

Regional Analysis: Photonic Neuron Chip Market

North America The North American photonic neuron chip market is driven by strong R&D investments in artificial intelligence and neuromorphic computing, particularly in the U.S. With tech giants like IBM and Intel pushing innovation boundaries, the region accounted for over 35% of global market revenue in 2024. Government initiatives such as the National Science Foundation’s $220 million AI Research Institutes program are accelerating adoption. However, high development costs and limited commercial-scale manufacturing capabilities currently constrain mass-market penetration. The telecommunications sector emerges as a key adopter, leveraging photonic neural networks for ultra-fast signal processing in 5G infrastructure.

Europe Europe maintains a technology-first approach through Horizon Europe’s €95.5 billion research program, with Germany and France leading in optical computing applications. Strict data protection regulations (GDPR) create unique demand for photonic chips’ inherent security advantages in data centers. The automotive sector shows promising adoption rates, with Continental and BMW investing in optical neural networks for autonomous driving systems. Challenges include fragmentation across national innovation policies and comparatively lower semiconductor fabrication capacity than Asia. The region’s emphasis on edge computing and IoT applications positions photonic neuron chips as strategic components for future smart infrastructure.

Asia-Pacific As the fastest growing market (projected 32.1% CAGR through 2032), Asia-Pacific benefits from China’s semiconductor self-sufficiency drive and Japan’s longstanding photonics expertise. China’s $1.4 trillion Digital Silk Road initiative includes substantial photonic computing investments, while South Korea’s Samsung leads in memory-integrated photonic architectures. Cost-competitive manufacturing coexists with cutting-edge research, though intellectual property concerns occasionally hinder international collaboration. India emerges as a dark horse, with 47% growth in AI startup funding creating unexpected demand for neural processing units. The region’s scale advantages make it poised to dominate production volumes by 2027.

South America Market development remains uneven, with Brazil accounting for 68% of regional photonic chip adoption. While academic partnerships with European institutions foster research capabilities, commercial deployment lags due to unstable technology budgets and reliance on imports. Bright spots include Chile’s astronomy-focused optical computing initiatives and Argentina’s healthcare AI applications. The lack of domestic semiconductor ecosystems forces reliance on foreign suppliers, creating price sensitivity that slows adoption of premium photonic solutions. Nevertheless, growing smart city projects in major urban centers suggest long-term potential as infrastructure modernizes.

Middle East & Africa The UAE and Saudi Arabia drive regional growth through sovereign wealth fund investments in AI infrastructure (exceeding $2.3 billion since 2020). While photonic adoption remains niche, purpose-built applications in oil/gas predictive maintenance and financial services show promise. Israel’s thriving deep tech scene produces innovative photonic startups, though scale challenges persist. Infrastructure gaps elsewhere in Africa limit market potential, despite growing recognition of photonic computing’s advantages for leapfrog technologies. Collaborative research centers like the African Photonics Initiative aim to build regional capacity, but progress depends on sustained funding commitments.

List of Key Photonic Neuron Chip Companies Profiled

Intel Corp (U.S.)

IBM Corp (U.S.)

BrainChip Holdings (Australia)

Samsung Group (South Korea)

Hewlett Packard Enterprise (U.S.)

General Vision (Switzerland)

Applied Brain Research (Canada)

The photonic neuron chip market is experiencing significant growth due to breakthroughs in artificial intelligence (AI) and neuromorphic computing. These chips, which mimic biological neurons using light rather than electricity, offer substantially higher speeds and energy efficiency compared to traditional silicon-based processors. The global AI hardware market is projected to grow at over 30% annually through 2030, creating strong demand for photonic neural networks. Recent developments in optical computing architectures have enabled photonic neuron chips to process complex neural networks with latency reductions up to 90% compared to electronic counterparts.

With data centers consuming approximately 200 terawatt-hours annually worldwide – comparable to some medium-sized countries’ total energy consumption – the need for energy-efficient computing has never been greater. Photonic neuron chips address this challenge directly, demonstrating power efficiency improvements up to 100x over conventional processors for AI workloads. Major cloud service providers are actively investing in photonic computing research to reduce their carbon footprints while maintaining computational performance. For instance, a recent industry report showed photonic AI accelerators could reduce data center energy consumption by 30-50% when deployed at scale.

The defense sector represents a key growth area for photonic neuron chips due to their inherent advantages in harsh environments. Unlike electronic systems, photonic processors are immune to electromagnetic interference and can operate across extreme temperature ranges. Global defense spending on AI and advanced computing technologies reached $30 billion in 2023, with applications including real-time signal processing, autonomous systems, and secure communications. Photonic neural networks enable critical capabilities such as jamming-resistant radar processing and high-speed battlefield decision making without the power constraints of traditional electronics.

The proliferation of edge computing and IoT devices presents significant opportunities for photonic neuron chips. Their low power consumption makes them ideal for battery-operated devices requiring advanced AI capabilities. Projections indicate the edge AI processor market will exceed $50 billion by 2030, with photonic solutions well-positioned to capture share in applications like smart sensors, autonomous drones, and real-time video analytics. Recent prototypes have demonstrated photonic neural networks consuming milliwatts of power while performing complex image classification tasks impossible for conventional microcontrollers.

Photonic neuron chips show strong potential as interfaces between quantum and classical computing systems. As quantum computing matures, hybrid quantum-photonic architectures could enable new classes of algorithms combining the strengths of both technologies. Early research indicates photonic neural networks may help mitigate quantum noise and assist in error correction, addressing key challenges in practical quantum computing. Several national quantum initiatives now include photonic neural networks as a research priority, signaling long-term government support for this convergence.

The medical field represents another high-potential market for photonic neuron chips. Their ability to process complex biological signals in real time could revolutionize applications from medical imaging to brain-machine interfaces. Recent advances have demonstrated photonic neural networks analyzing neural activity patterns for seizure prediction and processing optical coherence tomography data for early disease detection. With global healthcare AI investment surpassing $15 billion annually, photonic solutions could capture a growing portion of this market by offering superior performance for latency-sensitive medical applications.

The market is highly fragmented, with a mix of global and regional players competing for market share. To Learn More About the Global Trends Impacting the Future of Top 10 Companies https://semiconductorinsight.com/download-sample-report/?product_id=103240

Key Questions Answered by the Photonic Neuron Chip Market Report:

What is the current market size of Global Photonic Neuron Chip Market?

Which key companies operate in Global Photonic Neuron Chip Market?

What are the key growth drivers?

Which region dominates the market?

What are the emerging trends?

Browse More Reports:                                                     

CONTACT US:

City vista, 203A, Fountain Road, Ashoka Nagar, Kharadi, Pune, Maharashtra 411014

[+91 8087992013]

help@semiconductorinsight.com

[futuremarketreport]

Neuromorphic Computing Market Expected to Witness a Sustainable Growth over 2025 | Global Market Vision

The neuromorphic computing market is experiencing rapid growth, driven by the increasing demand for energy-efficient and real-time AI processing across various industries. This technology, inspired by the human brain's architecture, enables machines to process information more efficiently, making it ideal for applications in robotics, healthcare, automotive, and consumer electronics.​

Key trends shaping the market include the development of hybrid neuromorphic-conventional computing systems, advancements in edge AI and IoT, and the integration of neuromorphic computing with brain-computer interfaces. These innovations are expanding the potential applications of neuromorphic technology, from enhancing autonomous vehicle navigation to improving real-time data analysis in healthcare diagnostics.​G

Get Full PDF Sample Copy of Report (Including Full TOC, List of Tables & Figures, Chart) @ https://futuremarketconsulting.com/request-sample/53610

Key Market Players:

Brain Corporation

CEA-Leti

General Vision Inc.

Hewlett Packard Enterprise Development LP

HRL Laboratories, LLC

IBM

Intel Corporation

Knowm Inc.

Cognixion

BrainChip, Inc.

MindMaze

SAMSUNG

Vicarious

Bitbrain Technologies

Qualcomm Technologies, Inc.

Others

By Component (Hardware, Software, Services)

By Deployment (Edge, Cloud)

By Application (Signal Processing, Image Processing, Data Processing, Object Detection, Others)

By End-Use (Consumer Electronics, Automotive, Healthcare, Military & Defense, Others)

Key Target Audience:

• Neuromorphic Computing manufacturers and other stakeholders

• Organizations, forums and alliances related to Neuromorphic Computing distribution

• Government bodies such as regulating authorities and policy makers

• Market research organizations and consulting companies

The study is useful in providing answers to several critical questions that are important for industry stakeholders such as Neuromorphic Computing manufacturers, customers and policy makers. The study would also help them to target the growing segments over the coming years, thereby aiding the stakeholders in taking investment decisions and facilitating their expansion.

The following are the major objectives of the study.

To define, describe, and forecast the global Neuromorphic Computing market size on the basis of grade, application, type, and region

To provide detailed information regarding the significant factors influencing the growth of the market (drivers, restraints, opportunities, and industry-specific challenges)

To analyze the opportunities in the market for stakeholders and details of a competitive landscape for market leaders

To forecast the market size, in terms of value and volume, with respect to five main regions, namely, North America, Europe, Asia-Pacific, Latin America, and the Middle East & Africa

To strategically profile key players and comprehensively analyze their market shares and core competencies

To track and analyze competitive developments such as joint ventures, mergers & acquisitions, new product developments, and research & developments (R&D) in the Neuromorphic Computing market

During this research study, major players operating in the Neuromorphic Computing market in various regions have been identified, and their offerings, regional presence, and distribution channels have been analyzed through in-depth discussions. Top-down and bottom-up approaches have been used to determine the overall market size. Sizes of the other individual markets have been estimated using the percentage splits obtained through secondary sources such as Hoovers, Bloomberg BusinessWeek, and Factiva, along with primary respondents. The entire procedure includes the study of the annual and financial reports of the top market players and extensive interviews with industry experts such as CEOs, VPs, directors, and marketing executives for key insights (both qualitative and quantitative) pertaining to the market. The figure below shows the breakdown of the primaries on the basis of the company type, designation, and region considered during the research study.

Frequently asked questions

How much is the global Neuromorphic Computing market worth?

What was the value of the Neuromorphic Computing market in 2021?

At what CAGR is the Neuromorphic Computing market projected to grow in the forecast period (2022-2028)?

What is the leading segment in the market?

What is the key factor driving the market?

Which are the leading players in the market?

Which country held the highest market share in the market?

Which factors are expected to drive the adoption of the product?

Buy Exclusive Report @: https://futuremarketconsulting.com/buy-now/53610

NOTE: Our analysts monitoring the situation across the globe explains that the market will generate remunerative prospects for producers post the COVID-19 crisis. The report aims to provide an additional illustration of the latest scenario, economic slowdown, and COVID-19 impact on the overall industry.

Related Repots:

[gis56]

🧠💾 Brain-Inspired Chips? Neuromorphic Tech Is Growing FAST!

Neuromorphic semiconductor chips are revolutionizing AI hardware by mimicking the biological neural networks of the human brain, enabling ultra-efficient, low-power computing. Unlike traditional von Neumann architectures, these chips integrate spiking neural networks (SNNs) and event-driven processing, allowing real-time data analysis with minimal energy consumption. 

To Request Sample Report : https://www.globalinsightservices.com/request-sample/?id=GIS10673 &utm_source=SnehaPatil&utm_medium=Article

By leveraging advanced semiconductor materials, 3D chip stacking, and memristor-based architectures, neuromorphic chips significantly improve pattern recognition, autonomous decision-making, and edge AI capabilities. These advancements are critical for applications in robotics, IoT devices, autonomous vehicles, and real-time medical diagnostics, where low-latency, high-efficiency computing is essential. Companies like Intel (Loihi), IBM (TrueNorth), and BrainChip (Akida) are pioneering neuromorphic processors, paving the way for next-generation AI solutions that operate closer to biological cognition.

The integration of analog computing, in-memory processing, and non-volatile memory technologies enhances the scalability and performance of neuromorphic chips in complex environments. As the demand for edge AI, neuromorphic vision systems, and intelligent sensors grows, researchers are exploring synaptic plasticity, stochastic computing, and hybrid digital-analog designs to further optimize efficiency. These chips hold promise for neuromorphic supercomputing, human-machine interfaces, and brain-computer interfaces (BCIs), driving innovations in AI-driven healthcare, cybersecurity, and industrial automation. With the convergence of AI, semiconductor technology, and neuroscience, neuromorphic semiconductor chips will be the cornerstone of next-gen intelligent computing architectures, unlocking unprecedented levels of cognitive processing and energy-efficient AI.

#neuromorphiccomputing #aihardware #braininspiredcomputing #semiconductortechnology #spikingneuralnetworks #neuromorphicsystems #memristors #analogcomputing #intelligentprocessors #machinelearninghardware #edgedevices #autonomoussystems #eventdrivenprocessing #neuralnetworks #biomimeticai #robotics #aiattheneuromorphicedge #neuromorphicvision #chipdesign #siliconneurons #futurecomputing #hpc #smartai #inmemorycomputing #lowpowerai #bci #nextgenai #deeptech #cybersecurityai #intelligentsensors #syntheticintelligence #artificialcognition #computervision #braincomputerinterfaces #aiinnovation

[jobrxiv]

Artificial Synaptic Floating Gate Transistors Based on Hybrid Organic-Inorg IMEC Design & synthesis of novel hybrid organic-inorganic perovskites to boost the performance of floating-gate transistors toward neuromorphic computing a See the full job description on jobRxiv: https://jobrxiv.org/job/imec-27778-artificial-synaptic-floating-gate-transistors-based-on-hybrid-organic-inorg/?feed_id=89083 #chemistry #organic_synthesis #ScienceJobs #hiring #research

[ultrimio]

Teler Paper(v0)

Teler: version V0 (compiled in dec 2023 and subject to changes)

In the context of informatics, Teler refers to a new class of data-demodulation capable operating devices, that ideally transform sequential data(e.g:Python commands) through a specially constructed cognitive network, potentially capable of replacing entirely or working in hybrid configuration with everyday computers to perform various jobs but much faster than a simple personal computer.

Background While Transformers are the most widely used and documented architecture available in neural networks for especially chatbots and vision tasks, other interesting architectures appeared in the past. One of those is the Echo State Network(ESN), created by Jaeger in 2001. The ESN presents an astoundigly big reservoir of random data in the form of spiking neural oscillators constituing a "base" signal. An output is reconstructed by a trained readout operating on that signal. However that ESN is great for signals manipulation alone.

Part1: Teleoperations Teleoperating in informatics refers to the fact the Teler system functioning is related to various intelligent operations executed via exchange of signals between intelligent and fixed data "stations" directly on neuromorphic hardware. Stations is an improved base signal. That stations network is implemented on special hardware furnished with spiking neurons ideally operating in Thz range in photonics configurations, and retaining extraordinarily efficiency while making it suitable for edge local inference. That system thus can also be operated on a mix of resistive memristors, silicon neurons, analog spiking devices and etc.. but at various efficiencies. So in terms of informatics operations, Teler could be the next great thing because the main theory behind the functioning of it is a completely renovated and remodeled reservoir computing network to extend the characteristics of it. In fact that reservoir now rather act as a neural "capacity" of a multitude data stations instead of mere loosely coupled oscillators (which is technically simple but primitive). Now in the neural capacity technically each one of the station could send required data to the output unit while the involved math operations are being executed on a special readout called "tuner", required to display the solution. Not only, the architecture is also capable of actually executing high-level programming concepts directly in blazing fast way inheriting and extending the capabilities of differentiable neural computers theory. As an example executing directly BASIC commands could be made without wasting loops time in various hardware layers to bring data in and out of transistors as on Von Neumann computers. In this case (Teler) possess only a single layer, the layer 1, and that operates on layer 0 (physical) so the actual photonics hardware brings out the result immediately, making the Teler system capable of accelerating parallel tasks code execution beyond immagination and superfast compared the commonly used CPUs and GPUs.

Part2: is coming Attention is the new memory. In Teler cassettes are a type of attention… Particularities of the Teler TBD Particularities of the Cassette architecture TBD

[ophiocordyceps]

ok i need to know how v2 preserves itself and what kind of being it becomes in a sense, how it changes and what it's able to retain as it saves itself from its decay. ALSO very interested in gabriel's response to it - both like. how he takes the plans it enacts with regard to its study of angels and its transition into a new entity (hope it's not too much but i love these ideas uuwahhhh)

im not yet sure how gabe would react other than it being generally bittersweet !! (unless @presidentbungus has any input there), though he probably wouldnt've been too surprised when it first started studying angels since V2 was just extremely curious all around and initially it just wanted to learn more about them, since that was wayyy before any of end-of-life plans had ever even been relevant. but the whole backup thing is kind of stressful for V2, at least in a 'you need to get this done while you still can' sense. the whole reason phoenix is more of another, separate entity than just V2 resurrected is that I kind of took from the idea of how artificial intelligence works in the mass effect universe, where you can't even transfer an AI to different hardware without fundamentally altering it by virtue of slight variables between different bits of hardware, even when it's presumably identical on the surface, combined with the fact that in my interpretation the Vs computers are neuromorphic and modelled after the human brain and have artificial synapses and everything, so instead it decides to create some sort of esoteric hybrid processor (which takes a lot of trial and error and probably some suspension of disbelief) that'll last for a very, very long time (it's capable of regenerating itself, but puts a lot of strain on the rest of its body's systems, meaning phoenix has to do a lot more frequent upkeep) and then uploading some of it's memories and other archived data to it that it deems important or vital; mostly important events in its life and various bits of info that would otherwise be lost, and then it leaves gabriel the instructions for how to actually swap out the parts and upload the data backup onto it. it really is a sort of bittersweet 'came back wrong' sort of deal because of the inherent differences between iterations, which is something they would need to work out for sure. it's aware that it's a derivative of someone else who it can never truly be, but it doesn't dwell on that because it knows it was created out of it's predecessor's love for it's friend and to make sure someone would be there still with him, it has those memories and has at least part of that connection still.

phoenix is also initially just kind of. off-putting, vaguely. being an angelic machine (or a mechanical angel), it operates at far higher temperatures and its lights are all bleached white and are brighter and trying to output light from any possible place on its body. it's aware that it's not quite the same entity as V2, though it will respond to both that name and it's given designation (PHNX-AM or just phoenix, V2 named it that both to play into the idea of something being reborn from ashes and also riffing off its mechanical nature and angels as messengers). it's generally more hollow and blank seeming at first, though it does start to relearn a few of V2's mannerisms, and as time goes on it eventually begins modifying its body again into something more unique over the course of what would be several lifetimes both for optimization reasons and identity realization reasons. the whole dual nature thing probably messes with gabe for a good while at first, reminding him of things he's lost or left behind

instead of asking whether a doctor has a doctor, we're asking if angels can have guardian angels, basically.

[ai-briefing]

Neuromorphic Device with Low Power Consumption

Neuromorphic Device with Low Power Consumption

Compact, low–latency, and low–power computer systems are required for real–world sensory–processing applications. Hybrid memristive CMOS neuromorphic architectures, with their in–memory event–driven computing capabilities, present an appropriate hardware substrate for such tasks. To demonstrate the full potential of such systems and drawing inspiration from the barn owl’s neuroanatomy, CEA–Leti…

[techsciresearch]

Saudi Arabia Telecom ICT Infrastructure Market to grow at a CAGR of 9.01% until 2030 – TechSci Research

Increasing demand for cloud computing services, commercialization of 5G and growing mobile device penetration in the country are expected to drive the growth of Saudi Arabia Telecom ICT Infrastructure Market.

 According to TechSci Research report, “Saudi Arabia Telecom ICT Infrastructure Market By Component (Hardware, Software, Services), By Deployment Mode (On-Premises  v/s Cloud), By Network Type (Wireless v/s Fixed), By End User (Telecom Service Providers, Government, Enterprises, Others), By Region, Competition Forecast & Opportunities, 2016-2030F”, the Saudi Arabia Telecom ICT Infrastructure Market is forecast to grow at a rate of 9.01% to reach USD3631.16 million by 2030. Number of online transactions in telecom sector is growing rapidly in Saudi Arabia along with the increasing number of mobile devices. Telecom ICT infrastructure helps in billing & revenue management, service fulfillment, etc. Thus, owing to increasing online transaction and mobile devices, the country’s telecom ICT infrastructure market is also growing at a significant rate.

Adoption of cloud technologies is increasing due to high scalability, cost efficiency coupled with the ease of increasing need for data storage. Due to the aforementioned factors, telecom companies are adopting cloud-based solutions at a very rapid pace, which in turn, is driving the market of Telecom ICT Infrastructure in Saudi Arabia.

Browse 52 Figures spread through 127 Pages and an in-depth TOC on

" Saudi Arabia Telecom ICT Infrastructure Market "

https://www.techsciresearch.com/report/saudi-arabia-telecom-ict-infrastructure-market/4581.html

The Saudi Arabia telecom ICT infrastructure market is segmented based on component, deployment mode, network type, end user, company and region. Based on component, the market can be categorized into hardware, software and services. Services is the fastest growing segment by component in the Saudi Arabia Telecom ICT Infrastructure Market owing to the adoption of cloud technology throughout the country and increasing spending by the government in the telecom sector. Saudi Arabia aims to increase non-oil government revenue from USD43.5 billion to USD267 billion by 2030.

The hardware segment can further be segmented into networking equipment, transmitting device and end point device. The networking equipment hardware component is expected to hold a significant share in the market owing to the increasing proliferation of high speed mobile data and launch of 5G in the kingdom.

Central region dominates the Saudi Arabia Telecom ICT Infrastructure Market with a share of 31.84% in 2020 and is expected to dominate the market in the forecast period due to the presence of many telecom service providers in the central region including Riyadh. Moreover, increasing eGovernment services and growing population in the region is driving the demand for telecom infrastructure.

The software segment can further be segmented into Operations Support Software, Business Support Software, Network Management Software and Others. Operations Support Software (OSS) is the dominating software segment in the Saudi Arabia Telecom ICT Infrastructure Market. Operations support software helps in maintaining the network uptime and reducing the network downtime. It aids in service assurance which ensures customer retention, one of the top priorities among the telecom service providers due to stiff competition. Saudi companies are collaborating with MNCs to transform operations to a predictive and proactive automated operation.

Based on services, the market can further be segmented in Professional and Managed services. Managed services are on-going and can cover many aspects of an organization’s IT needs on a daily basis whereas professional services are project-based and usually address a specific problem or challenge. Professional services help businesses overcome specific challenges through a dedicated project, such as a cloud migration or the deployment of new hardware.

ZTE Corporation, Ciena Corporation, Cisco Systems, Inc, CommScope, Inc, Fortinet, Inc, Fujitsu Ltd, Huawei Technologies Co., Ltd, Juniper Networks Inc, Saudi Telecom Company, Etihad Etisalat Co. (Mobily) and others are some of the leading players operating in Saudi Arabia telecom ICT infrastructure market. Companies operating in the market are using organic strategies, partnerships and collaborations to boost their shares in the Saudi Arabia Telecom ICT Infrastructure Market. Significant players in the market are concentrating on accomplishing ideal operational expenses, upgrading the system efficiency, enhancing precision in responses, boosting productivity with high funding in R&D, and merging with small players to support the competitive Saudi Arabia Telecom ICT Infrastructure Market.

Download Sample Report  @ https://www.techsciresearch.com/sample-report.aspx?cid=4581  

Customers can also request for 10% free customization on this report.

 “Over the last few years, the deployment of on-premises Operational Support systems (OSS) and Business Support Systems (BSS) solutions dominated the market. However, cloud deployment is becoming the favorable deployment type among telecom service providers, owing to which this segment is expected to overtake on-premises deployment in the coming years. Factors like stiff competition, quickly changing customer demand for new services and pressure to reduce price are forcing the service providers to go for cloud deployment as it reduces the initial cost of on-premises deployment. Cloud deployment also helps service providers to get infrastructure as per their requirements.” said Mr. Karan Chechi, Research Director with TechSci Research, a research based global management consulting firm.

According to TechSci Research “Saudi Arabia Telecom ICT Infrastructure Market By Component (Hardware, Software, Services), By Deployment Mode (On-Premises v/s Cloud), By Network Type (Wireless v/s Fixed), By End User (Telecom Service Providers,  Government, Enterprises, Others), By Region, Competition Forecast & Opportunities, 2016-2030F” has evaluated the future growth potential of Telecom infrastructure in the Saudi Arabia Market and provides statistics and information on market structure, size, share and future growth. The report is intended to provide cutting-edge market intelligence and help decision makers take sound investment decisions. Besides, the report also identifies and analyzes the emerging trends along with essential drivers, challenges and opportunities present in Saudi Arabia Telecom ICT Infrastructure Market.”

Browse Related Reports

Vietnam Cloud Storage Market By Component (Solution v/s Service), By Deployment Mode (Public, Private, Hybrid), By Organization Size (Large Enterprises v/s SMEs), By End User Industry (IT & Telecommunications, BFSI, Healthcare, Retail & E-Commerce, Manufacturing & Engineering, Others), By Company, By Region, Forecast & Opportunities, 2026 2026

https://www.techsciresearch.com/report/vietnam-cloud-storage-market/1991.html

Global Neuromorphic Computing Market By Offering (Hardware, Software) By Deployment (Edge Computing, Deploy Computing) By Application (Image Recognition, Signal Recognition, Data Mining) By Vertical (Aerospace, Military, Defense, Automotive, Consumer Electronics, Industrial, Medical, IT & Telecommunication, Others), By Region, Company Forecast & Opportunities, 2026

https://www.techsciresearch.com/report/neuromorphic-computing-market/7592.html

United States Warehouse Robotics Market, By Type (Mobile Robots, Articulated Robots, Cylindrical Robots, Scara Robots, Parallel Robots & Cartesian Robots), By Software (Warehouse Management System, Warehouse Control System & Warehouse Execution Systems), By Payload (0.5 Kg to 10 Kg, 11 Kg to 80 Kg, 81 Kg to 180 Kg, 181 Kg to 300 Kg, 301 Kg to 900 Kg ad More than 900 Kg), By End-User (E-Commerce, Automotive, Electricals and Electronics, Chemical, Rubber & Plastics, Food and Beverages, Pharmaceutical & Others),Competition Forecast & Opportunities, 2016 – 2026

 https://www.techsciresearch.com/report/united-states-warehouse-robotics-market/2173.html

About TechSci Research

TechSci Research is a leading global market research firm publishing premium market research reports. Serving 700 global clients with more than 600 premium market research studies, TechSci Research is serving clients across 11 different industrial verticals. TechSci Research specializes in research-based consulting assignments in high growth and emerging markets, leading technologies and niche applications. Our workforce of more than 100 full-time Analysts and Consultants employing innovative research solutions and tracking global and country-specific high growth markets helps TechSci clients to lead rather than follow market trends.

Contact

Mr. Ken Mathews

708 Third Avenue,

Manhattan, NY,

New York – 10017

Tel: +1-646-360-1656

Email: sales@techsciresearch.com

[damiencordle]

Joshua Damien Cordle. I Found This Interesting

'Egg carton' quantum dot array could lead to ultralow power devices

A new path toward sending and receiving information with single photons of light has been discovered by an international team of researchers led by the University of Michigan.

Their experiment demonstrated the possibility of using an effect known as nonlinearity to modify and detect extremely weak light signals, taking advantage of distinct changes to a quantum system to advance next generation computing.

Today, as silicon-electronics-based information technology becomes increasingly throttled by heating and energy consumption, nonlinear optics is under intense investigation as a potential solution. The quantum egg carton captures and releases photons, supporting "excited" quantum states while it possesses the extra energy. As the energy in the system rises, it takes a bigger jump in energy to get to that next excited state -- that's the nonlinearity.

"Researchers have wondered whether detectable nonlinear effects can be sustained at extremely low power levels -- down to individual photons. This would bring us to the fundamental lower limit of power consumption in information processing," said Hui Deng, professor of physics and senior author of the paper in Nature.

"We demonstrated a new type of hybrid state to bring us to that regime, linking light and matter through an array of quantum dots," she added.

The physicists and engineers used a new kind of semiconductor to create quantum dots arranged like an egg carton. Quantum dots are essentially tiny structures that can isolate and confine individual quantum particles, such as electrons and other, stranger things. These dots are the pockets in the egg carton. In this case, they confine excitons, quasi-particles made up of an electron and a "hole." A hole appears when an electron in a semiconductor is kicked into a higher energy band, leaving a positive charge behind in its usual spot. If the hole shadows the electron in its parallel energy band, the two are considered as a single entity, an exciton.

In conventional devices -- with little to no nonlinearity -- the excitons roam freely and scarcely meet with each other. These materials can contain many identical excitons at the same time without researchers noticing any change to the material properties.

However, if the exciton is confined to a quantum dot, it becomes impossible to put in a second identical exciton in the same pocket. You'll need an exciton with a higher energy if you want to get another one in there, which means you'll need a higher energy photon to make it. This is known as quantum blockade, and it's the cause of the nonlinearity.

But typical quantum dots are only a few atoms across -- they aren't on a usable scale. As a solution, Deng's team created an array of quantum dots that contribute to the nonlinearity all at once.

The team produced this egg carton energy landscape with two flakes of semiconductor, which are considered two-dimensional materials because they are made of a single molecular layer, just a few atoms thick. 2D semiconductors have quantum properties that are very different from larger chunks. One flake was tungsten disulfide and the other was molybdenum diselenide. Laid with an angle of about 56.5 degrees between their atomic lattices, the two intertwined electronic structures created a larger electronic lattice, with pockets about 10 atoms across.

In order for the array of quantum dots inside the 2D semiconductor to be controlled as a group with light, the team built a resonator by making one mirror at the bottom, laying the semiconductor on top of it, and then depositing a second mirror on top of the semiconductor.

"You need to control the thickness very tightly so that the semiconductor is at the maximum of the optical field," said Zhang Long, a postdoctoral research fellow in the Deng lab and first author on the paper.

With the quantum egg carton embedded in the mirrored "cavity" that enabled red laser light to resonate, the team observed the formation of another quantum state, called a polariton. Polaritons are a hybrid of the excitons and the light in the cavity. This confirmed all the quantum dots interact with light in concert. In this system, Deng's team showed that putting a few excitons into the carton led to a measurable change of the polariton's energy -- demonstrating nonlinearity and showing that quantum blockade was occurring.

"Engineers can use that nonlinearity to discern energy deposited into the system, potentially down to that of a single photon, which makes the system promising as an ultra-low energy switch," Deng said.

Switches are among the devices needed to achieve ultralow power computing, and they can be built into more complex gates.

"Professor Deng's research describes how polariton nonlinearities can be tailored to consume less energy," said Michael Gerhold, program manager at the Army Research Office, an element of the U.S. Army Combat Capabilities Development Command's Army Research Laboratory. "Control of polaritons is aimed at future integrated photonics used for ultra-low energy computing and information processing that could be used for neuromorphic processing for vision systems, natural language processing or autonomous robots."

The quantum blockade also means a similar system could possibly be used for qubits, the building blocks for quantum information processing. One forward path is figuring out how to address each quantum dot in the array as an individual qubit. Another way would be to achieve polariton blockade, similar to the exciton blockade seen here. In this version, the array of excitons, resonating in time with the light wave, would be the qubit.

Used in these ways, the new 2D semiconductors have potential for bringing quantum devices up to room temperature, rather than the extreme cold of liquid nitrogen or liquid helium.

"We are coming to the end of Moore's Law," said Steve Forrest, the Peter A. Franken Distinguished University Professor of Electrical Engineering and co-author of the paper, referring to the trend of the density of transistors on a chip doubling every two years. "Two dimensional materials have many exciting electronic and optical properties that may, in fact, lead us to that land beyond silicon."

Story Source:

Materials provided by University of Michigan. Original written by Kate McAlpine. Note: Content may be edited for style and length.

Journal Reference:

Long Zhang, Fengcheng Wu, Shaocong Hou, Zhe Zhang, Yu-Hsun Chou, Kenji Watanabe, Takashi Taniguchi, Stephen R. Forrest, Hui Deng. Van der Waals heterostructure polaritons with moiré-induced nonlinearity. Nature, 2021; 591 (7848): 61 DOI: 10.1038/s41586-021-03228-5

[renatoferreiradasilva]

\documentclass[11pt]{article} \usepackage{amsmath, amssymb, amsfonts} \usepackage{geometry} \usepackage{graphicx} \usepackage{hyperref} \geometry{margin=1in} \title{Spectral Foundations for Hybrid Optoelectronic Computing Architectures: From Riemann Zeros to Physical AI Hardware} \author{Renato Ferreira da Silva \ \texttt{renatoferreiradasilva1980@yahoo.com} \ ORCID: \href{https://orcid.org/0009-0003-8908-481X}{0009-0003-8908-481X}} \date{\today}

\begin{document}

\maketitle

\begin{abstract} This article establishes a theoretical bridge between spectral models of the Riemann zeta zeros and the design principles of hybrid optoelectronic computing architectures. By interpreting data flow and signal processing as spectral phenomena, we draw on Schr\"odinger operators with machine-learned potentials to inform the structural logic of neuromorphic chips. Our approach uses the Gaussian Unitary Ensemble (GUE) as a universal benchmark for both numerical accuracy and hardware robustness. We argue that such architectures are not merely computational devices, but physical realizations of spectral operators whose eigenvalues encode semantic and functional states. This spectral paradigm enables scalable, reconfigurable, and energetically efficient AI hardware that emulates fundamental structures from number theory and quantum mechanics. \end{abstract}

\section{Introduction} The convergence of high-performance computing, artificial intelligence, and photonics demands a foundational rethinking of how computation is physically realized. While Moore's Law slows, spectral methods rooted in mathematical physics offer a new direction. In particular, the Hilbert--P\'olya conjecture---which suggests that the nontrivial zeros of the Riemann zeta function correspond to the eigenvalues of a self-adjoint operator---invites a reinterpretation of computation as a spectral phenomenon. This work connects spectral operator modeling with hardware design, offering a principled framework for constructing optoelectronic systems whose logic is derived from eigenvalue dynamics.

\section{Spectral Operator Framework} We consider operators of the form: [ \mathcal{L} = -\frac{d^2}{dx^2} + V(x), ] with Dirichlet boundary conditions and potentials of the form: [ V(x) = \sum_{n=0}^K c_n H_n(x) e^{-x^2/2}, ] where $H_n(x)$ are Hermite polynomials and $c_n$ are trainable parameters. The eigenvalues of $\mathcal{L}$, computed numerically, align with the statistical properties of the nontrivial zeros of the Riemann zeta function. We leverage neural networks to learn mappings $c_n \mapsto \lambda_j$, interpreting these as functional transformations from hardware configurations to observable spectral states.

\section{GUE as Design Metric} The Gaussian Unitary Ensemble provides a statistical benchmark for spectral rigidity and eigenvalue spacing. In our context, it serves dual roles: \begin{itemize} \item In simulation: to validate the fidelity of the learned operator. \item In hardware: to ensure robust optical communication, minimal crosstalk, and consistent signal coherence across filaments. \end{itemize} Designing hardware whose delay paths, optical resonances, and thermal fluctuations approximate GUE behavior results in physical stability under computational load.

\section{Hardware Realization of Spectral Paradigms} Each computational unit (UC) in the hybrid optoelectronic architecture is interpreted as a physical realization of a node in a spectral graph: \begin{itemize} \item Optical interconnects act as eigenmode couplings. \item Modulation depths and frequencies correspond to potential configurations. \item Thermal gradients and electromagnetic fields define boundary conditions. \end{itemize} Hence, computation is no longer symbolic but spectral---defined by eigenvalue distributions and dynamic wave propagation.

\section{Spectral Learning and Inference} Using physics-informed neural networks (PINNs), we train the system to infer potential shapes from target spectral outputs. This establishes a feedback loop where: \begin{enumerate} \item Input data modulates physical parameters. \item Spectral output is measured optically. \item Machine learning adjusts the configuration for optimal response. \end{enumerate} This feedback aligns with biological principles of neural plasticity, offering a pathway to hardware-level learning.

\section{Conclusion and Outlook} We propose a new paradigm in AI hardware design rooted in spectral operator theory. By grounding hardware architectures in the spectral behavior of Schr\"odinger-type systems, validated through GUE statistics and machine learning, we lay the foundation for scalable, efficient, and intelligent photonic computing. Future work will explore: \begin{itemize} \item Topological extensions to noncommutative geometries. \item Quantum analogs with entangled photonic states. \item Integration with health diagnostics and real-time physical simulation. \end{itemize}

\end{document}

[tastydregs]

How Giving Robots a Hybrid, Human-Like ‘Brain’ Can Make Them Smarter

Squeezing a lot of computing power into robots without using up too much space or energy is a constant battle for their designers. But a new approach that mimics the structure of the human brain could provide a workaround.

The capabilities of most of today’s mobile robots are fairly rudimentary, but giving them the smarts to do their jobs is still a serious challenge. Controlling a body in a dynamic environment takes a surprising amount of processing power, which requires both real estate for chips and considerable amounts of energy to power them.

As robots get more complex and capable, those demands are only going to increase. Today’s most powerful AI systems run in massive data centers across far more chips than can realistically fit inside a machine on the move. And the slow death of Moore’s Law suggests we can’t rely on conventional processors getting significantly more efficient or compact anytime soon.

That prompted a team from the University of Southern California to resurrect an idea from more than 40 years ago: mimicking the human brain’s division of labor between two complimentary structures. While the cerebrum is responsible for higher cognitive functions like vision, hearing, and thinking, the cerebellum integrates sensory data and governs movement, balance, and posture.

When the idea was first proposed the technology didn’t exist to make it a reality, but in a paper recently published in Science Robotics, the researchers describe a hybrid system that combines analog circuits that control motion and digital circuits that govern perception and decision-making in an inverted pendulum robot.

“Through this cooperation of the cerebrum and the cerebellum, the robot can conduct multiple tasks simultaneously with a much shorter latency and lower power consumption,” write the researchers.

The type of robot the researchers were experimenting with looks essentially like a pole balancing on a pair of wheels. They have a broad range of applications, from hoverboards to warehouse logistics—Boston Dynamics’ recently-unveiled Handle robot operates on the same principles. Keeping them stable is notoriously tough, but the new approach managed to significantly improve all digital control approaches by radically improving the speed and efficiency of computations.

Key to bringing the idea alive was the recent emergence of memristors—electrical components whose resistance relies on previous input, which allows them to combine computing and memory in one place in a way similar to how biological neurons operate.

The researchers used memristors to build an analog circuit that runs an algorithm responsible for integrating data from the robot’s accelerometer and gyroscope, which is crucial for detecting the angle and velocity of its body, and another that controls its motion. One key advantage of this setup is that the signals from the sensors are analog, so it does away with the need for extra circuitry to convert them into digital signals, saving both space and power.

More importantly, though, the analog system is an order of magnitude faster and more energy-efficient than a standard all-digital system, the authors report. This not only lets them slash the power requirements, but also lets them cut the processing loop from 3,000 microseconds to just 6. That significantly improves the robot’s stability, with it taking just one second to settle into a steady state compared to more than three seconds using the digital-only platform.

At the minute this is just a proof of concept. The robot the researchers have built is small and rudimentary, and the algorithms being run on the analog circuit are fairly basic. But the principle is a promising one, and there is currently a huge amount of R&D going into neuromorphic and memristor-based analog computing hardware.

As often turns out to be the case, it seems like we can’t go too far wrong by mimicking the best model of computation we have found so far: our own brains.

Image Credit: Photos Hobby / Unsplash

[appliancesreviews]

HiSilicon Kirin SoC in Huawei and Honor smartphones Review

Of course, the chipset is one of the main components of a modern smartphone, affecting its performance. Therefore, many choose a smartphone taking into account the capabilities of chipset. Accordingly, all smartphone manufacturers pay great attention to this aspect. As known, System-on-a-Chip (SoC) is an integrated circuit that integrates all components of electronic system. Typically, a SoC contains one or more DSP (Digital Signal Processing) cores. But almost all modern smartphones use multiprocessor systems (MPSoC). In addition, a modern SoC often includes a GPU, memory controller, GSM, 3G and 4G LTE modules, fragments of GPS, USB, NFC, Bluetooth, and cameras. The list of leaders in SoC-platforms segment includes the American Qualcomm Snapdragon (Qualcomm Inc), Nvidia Tegra (NVIDIA Corporation), TI OMAP (Texas Instruments), Chinese MediaTek (MediaTek Inc). In addition, some large companies are developing their own chipsets. For example, this list includes ST-Ericsson NovaThor (Sony), Samsung Exynos (Samsung), Apple Ax (iPhone, Apple Inc), Kirin (Huawei Honor, HiSilicon Technologies Co., Ltd). As known, in 1991, Huawei formed the HiSilicon Technologies division. In 2004, it became an independent company and began to create its own RISC processor based on a license from the British ARM. Today, HiSilicon Technologies Co Ltd is developing high-performance mobile processors, successfully competing even with Qualcomm Snapdragon. The list of their latest chipsets of 2018 and 2019 includes the Kirin 970, 980, 990, 990 5G and Kirin 810.

HiSilicon Kirin 970

In Berlin at IFA 2017, Huawei introduced the flagship single-crystal HiSilicon Kirin 970 chipset. According to the company, Kirin 970 is the first Neural Processing Unit (NPU) or AI accelerator. As known, NPUs use the cluster asynchronous architecture developed at Cornell University. Unlike traditional computing architectures, this highly specialized logic is designed to create different types of artificial neural networks. Usually, each core contains a task scheduler, its own memory such as SRAM and a router for communication with other kernels. One chip can contain several thousand cores. The manufacture of a single-chip Kirin 970 system uses 10nm technology with the placement of 5.5 billion transistors in an area of about a square centimeter. The configuration of this SoC includes four ARM Cortex-A73 cores (2.4 GHz) and four ARM Cortex-A53 cores (1.8 GHz), combined according to the standard big.LITTLE scheme. Additionally, the Kirin 970 for the first time contains the 12-core Mali G72MP12 GPU (ARM GPU). Moreover, the built-in LTE Advanced Pro (4.5G) modem supports up to 1.2 Gb/s, and a dual DSP provides motion capture, face detection, four-level hybrid autofocus and improved shooting of moving objects even in low light. Additionally, Kirin 970 supports LPPDR 4X RAM (1883 MHz) and UFS 2.1, a 4K video codec, HDR10, a separate Security Engine with TEE and inSE support, an i7 sensor, and HiFi sound.

HiSilicon Kirin 980

A year later, at IFA 2018, Huawei introduced the new flagship Kirin 980. Kirin 980 became the first Huawei 7nm chipset. As known, it directly affects the energy consumption and, accordingly, the heating the processor during operation. SoC Kirin 980 has 6.9 billion transistors, which is 1.6 times more compared to Kirin 970. Its eight cores are divided into three clusters, including two Cortex-A76 with a frequency of 2.6 GHz, two Cortex-A76 with a frequency of 1.92 GHz and four Cortex-A55 with a frequency of 1.8 GHz. In addition, it uses the integrated GPU Mali-G76 MP10. According to the company, the performance increase of Kirin 980 vs 970 for the CPU reaches 58-75%, GPU - 46% with an increase in energy efficiency by 178%. In addition, the company doubled the number of NPU units to accelerate AI apps. In fact, the neuromorphic Huawei processor has become dual-core, capable of recognizing up to 4,500 images per minute. Additionally, the Kirin 980 uses the fourth generation ISP (image signal processor). The new SoC debuted in the flagship Mate 20 и Mate 20 Pro smartphones of 2018.

Kirin 810

In 2019, the company continued to improve its chipsets, introducing Kirin 810, 990 and 990 5G. In summer, the company introduced the new 7nm 8-core Kirin 810 with two ARM Cortex-A76 clocked up to 2.27 GHz and six ARM Cortex-A55 clocked up to 1.88 GHz. It also has an ARM Mali-G52 MP6 GPU and an AI task processing unit based on the new Huawei DaVinci architecture, which reduces energy consumption. Key Features: - 7nm technology; - the high quality of the built-in modems provides a very stable connection; - the improved Mali-G52 MP6 GPU significantly improves gaming performance, outperforming even the Adreno 618 (Qualcomm Snapdragon) in the GFXBench test; - the neural processor module on the Huawei DaVinci architecture in combination with the Huawei HiAi 2.0 platform surpasses even the Snapdragon 855 and Kirin 980 in the performance test for AI tasks.

Kirin 990 and Kirin 990 5G

At IFA 2019, the company introduced the new flagship Kirin 990 and Kirin 990 5G. This new 7nm SoC for the first time includes a 5G modem with support for NSA and SA networks. In addition, engineers were able to place 10.3 billion transistors on it due to the use of innovative EUV lithography, which forms the elements of electronic microcircuits smaller than 45 nm. Additionally, the new chipset contains a neuromorphic coprocessor with Da Vinchi architecture and a 16-core graphics core. Advanced adaptive receiver can work on any 5G NR (New Radio) networks, including Non-Standalone (NSA) and Standalone (SA). Today, companies are building 5G mainly to improve mobile broadband (eMBB) by increasing the speed of data transmission in the network. Today, it uses the 4G LTE infrastructure, operating in the millimeter range. Such networks are called non-autonomous. But this is only the first stage of 5G implementation. The next stage involves the creation of a new infrastructure with its own network core, which will additionally use low and medium frequencies. Of course, this will increase the coverage of 5G networks and the stability of high-speed data transmission. Such 5G networks will be called Standalone. Thus, Huawei phones already Standalone support. Additionally, Kirin 990 supports the use of two SIM cards (2G + 3G + 4G + 5G or 2G + 3G + 4G). Kirin 990 5G uses 2 + 2 + 4 configuration with two Cortex-A76 (2.86 GHz), two Cortex-A76 (2.36 GHz) and four Cortex A-55 (1.95 GHz). Kirin 990 is manufactured without the use of EUV lithography. It does not have a 5G modem, operates at lower frequencies, and its neuromorphic processor uses only one high-performance core.

Conclusion

Today we can note the rapid improvement of HiSilicon Kirin chipsets. In just 3 years, the company introduced 5 new chipsets, each of which justifiably claims the highest level. As a result, today HiSilicon Kirin successfully competes even with Qualcomm Snapdragon. For example, Kirin 990 5G shows the same performance as Qualcomm Snapdragon 855, but has fewer memory channels, which affects the speed of data exchange, does not have NX-bit (No-eXecute) and does not use HMP (heterogeneous multiprocessor). But the price / quality ratio compensates for these differences. Of course, Kirin chipsets to a large extent ensure the success of Huawei and Honor smartphones, complementing its new Harmony OS and excellent cameras manufactured in collaboration with the famous German Leica Camera AG. In addition, Harris Interactive conducted a study of the reliability of smartphones by analyzing 130,050 cases of after-sales appeals from customers of a large European Darty retail network. According to its results, Huawei and Apple have the lowest percentage of failures and breakdowns. Thus, Huawei, Honor and iPhone smartphones today can be considered the most reliable. Of course, all these factors retain excellent prospects for the smartphones of the Chinese giant. This video offers Speed, Gaming & Screen Test comparison of Huawei Mate 30 with Kirin 990 5G vs iPhone 11 with A13 Bionic. Read the full article

[cryptofeedzposts]

What’s in a name? The ‘deep learning’ debate

Monday’s historic debate between machine learning luminary Yoshua Bengio and machine learning critic Gary Marcus spilled over into a tit for tat between the two in the days following, mostly about the status of the term “deep learning.”

The history of the term deep learning shows that the use of it has been opportunistic at times but has had little to do in the way of advancing the science of artificial intelligence. Hence, the current debate will likely not go anywhere, ultimately. 

Monday night’s debate found Bengio and Marcus talking about similar-seeming end goals, things such as the need for “hybrid” models of intelligence, maybe combining neural networks with something like a “symbol” class of object. The details were where the two argued about definitions and terminology. 

In the days that followed, Marcus, in a post on Medium, observed that Bengio seemed to have white-washed his own recent critique of shortcomings in deep learning. And Bengio replied, in a letter on Google Docs linked from his Facebook account, that Marcus was presuming to tell the deep learning community how it can define its terms. Marcus responded in a follow-up post by suggesting the shifting descriptions of deep learning are “sloppy.” Bengio replied again late Friday on his Facebook page with a definition of deep learning as a goal, stating, “Deep learning is inspired by neural networks of the brain to build learning machines which discover rich and useful internal representations, computed as a composition of learned features and functions.” Bengio noted the definition did not cover the “how” of the matter, leaving it open. 

Also: Devil’s in the details in Historic AI debate

The term “deep learning” has emerged a bunch of times over the decades, and it has been used in different ways. It’s never been rigorous, and doubtless it will morph again, and at some point it may lose its utility. 

Jürgen Schmidhuber, who co-developed the “long-short term memory” form of neural network, has written that the AI scientist Rina Dechter first used the term “deep learning” in the 1980s. That use was different from today’s usage.  Dechter was writing about methods to search a graph of a problem, having nothing much to do with deep networks of artificial neurons. But there was a similarity: she was using the word “deep” as a way to indicate the degree of complexity of a problem and its solution, which is what others started doing in the new century.

The same kind of heuristic use of deep learning started to happen with Bengio and others around 2006, when Geoffrey Hinton offered up seminal work on neural networks with many more layers of computation than in past. Starting that year, Hinton and others in the  field began to refer to “deep networks” as opposed to earlier work that employed collections of just a small number of artificial neurons.

From Yoshua Bengio’s slides for the AI debate with Gary Marcus, December 23rd.

Yoshua Bengio

So deep learning emerged as a very rough, very broad way to distinguish a layering approach that makes things such as AlexNet work. 

In the meantime, as Marcus suggests, the term deep learning has been so successful in the popular literature that it has taken on a branding aspect, and it has become a kind-of catchall that can sometimes seem like it stands for anything. Marcus’s best work has been in pointing out how cavalierly and irresponsibly such terms are used (mostly by journalists and corporations), causing confusion among the public. Companies with “deep” in their name have certainly branded their achievements and earned hundreds of millions for it. So the topic of branding is in some sense unavoidable.

Bengio’s response implies he doesn’t much care about the semantic drift that the term has undergone because he’s focused on practicing science, not on defining terms. To him, deep learning is serviceable as a placeholder for a community of approaches and practices that evolve together over time. 

Also: Intel’s neuro guru slams deep learning: ‘it’s not actually learning’

Probably, deep learning as a term will at some point disappear from the scene, just as it and other terms have floated in and out of use over time. 

There was something else in Monday’s debate, actually, that was far more provocative than the branding issue, and it was Bengio’s insistence that everything in deep learning is united in some respect via the notion of optimization, typically optimization of an objective function. That could be a loss function, or an energy function, or something else, depending on the context. 

In fact, Bengio and colleagues have argued in a recent paper that the notion of objective functions should be extended to neuroscience. As they put it, “If things don’t ‘get better’ according to some metric, how can we refer to any phenotypic plasticity as ‘learning’ as opposed to just ‘changes’?”

That’s such a basic idea, it seems so self-evident, that it almost seems trivial for Bengio to insist on it. 

But it is not trivial. Insisting that a system optimizes along some vector is a position that not everyone agrees with. For example, Mike Davies, head of Intel’s “neuromorphic” chip effort, this past February criticized back-propagation, the main learning rule used to optimize in deep learning, during a talk at the International Solid State Circuits Conference.

Davies’s complaint is that back-prop is unlike human brain activity, arguing “it’s really an optimization procedure, it’s not actually learning.” 

Thus, deep learning’s adherents have at least one main  tenet that is very broad but also not without controversy.

The moral of the story is, there will always be something to argue about. 

[jobrxiv]

Artificial Synaptic Floating Gate Transistors Based on Hybrid Organic-Inorg IMEC Design & synthesis of novel hybrid organic-inorganic perovskites to boost the performance of floating-gate transistors toward neuromorphic computing a See the full job description on jobRxiv: https://jobrxiv.org/job/imec-27778-artificial-synaptic-floating-gate-transistors-based-on-hybrid-organic-inorg/?feed_id=86759 #chemistry #organic_synthesis #ScienceJobs #hiring #research