BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Tothemoon24
Top 20
Embedded World 2023: VDC’s “Embeddy” Award Winners
by Dan Mandell & Chris Rommel | 3/28/2023Last week, VDC Research attended the Embedded World 2023 conference in Nuremberg, Germany and held dozens of face-to-face meetings with exhibitors. Following a barren 2020 and relatively muted attendance last year, Embedded World 2023 was a clear indication that this industry is a small step away from back to normal conference and tradeshow attendance and activities. The halls were packed and bustling like times of old, with everyone sharing a sense of relief and rejuvenated appreciation for in-person events, meetings, and connections.
VDC created and named the Embeddy Awards to highlight companies announcing important advances in the IoT and embedded software, hardware, and services industries. The Embeddy is awarded for the most cutting-edge product or service available to embedded software developers and system engineers. Nominees were judged across specific criteria including corporate, technological, and industry significance, as well as the most cutting-edge hardware and software solutions or services.
The 2023 Embeddy Award Winners include:
- IP: BrainChip Akida 2nd Generation
- HARDWARE: Eurotech ReliaCOR Secure Edge AI
- SOFTWARE: 1NCE OS
Embeddy IP Winner: BrainChip
Akida 2nd Generation: Leading up to Embedded World 2023, BrainChip, a commercial producer of ultra-low power neuromorphic AI IP, launched the 2nd generation of its Akida platform to drive provide direct support for the embedded AI processor market, which is rapidly growing, particularly at the far edge. More complex networks like RESNET-50 can be processed by Akida without CPU intervention and the platform supports 8-bit weights to go with existing 4,2,1 bit support. Akida also adds an efficient vision transformer implementation for acceleration and can also handle 3D data for applications like video object detection and target tracking as well as 1D time series data for streaming data on devices with limited memory, battery, and compute resources. Akida features separable spatial temporal convolutions called Temporal Event Based Neural Nets that allow for simpler and efficient AI processing solutions.
Neuromorphia
fact collector
Interesting. I also didn't know that...Renesas counts major OEMs including everyone from Toyota to Tesla as key customers, with the former also being a shareholder in the company.A brief article on Renesas about how they formed and what they do. I wasn’t aware Toyota was a shareholder!
- In 2002, Mitsubishi Electric and Hitachi decided to consolidate their semiconductor businesses (excluding their DRAM, or dynamic random access memory business, which was merged separately) into a company called Renesas.
- In 2010, Renesas merged with the semiconductor operations of fellow Japanese company NEC Electronics to further expand its footprint, which brought about a minor name change to Renesas Electronics Corporation (from the former Renesas Technology Corporation).
- Renesas claims its R-Car V4H SoC is suitable for cars equipped with Level 3 autonomous driving features, and its maximum performance of 34 trillion operations per second facilitates activities such as high-speed image recognition and processing of objects that have been identified by surround-view cameras, radar and LiDAR systems.
- Renesas counts major OEMs including everyone from Toyota to Tesla as key customers, with the former also being a shareholder in the company
Inside the Suppliers: Renesas
The semiconductor shortage is crunching supply of new cars and forcing manufacturers to sell models with a pared-back feature set.www.carexpert.com.au
Renesas Electronics Corporation: Shareholders Board Members Managers and Company Profile | JP3164720009 | MarketScreener
Renesas Electronics Corporation: Company profile, business summary, shareholders, managers, financial ratings, industry, sector and market information | Japan Exchange: 6723 | Japan Exchange
www.marketscreener.com
Attachments
cassip
Regular
Today at "Ingenieur-News" a company was mentioned: "Denkweit":
denkweit.com
Batterie tests are one usecase.
DENKweit Offline Analyzer and Operator Software
Integrate our technology into your product or production. Use our modular offline analyzer or our APIs/DLLs
https://denkweit.com/en/denknetze-landing-page/denknetze-technology/
news from 08.03.2023:
www.pressebox.de
This is how Deep Learning can be used quickly, easily and cost-effectively - by everyone in all industries. IDS NXT is the all-in-one system for the use of intelligent cameras. You benefit from user-friendly workflows and perfectly coordinated hardware and software.
They are linked to edge-ai + vision alliance:
www.edge-ai-vision.com
DENKweit GmbH
DENKweit specializes in AI-based image analysis and magnetic field imaging. Our solutions enable companies to use complex technologies simply and profitably.
Batterie tests are one usecase.
Image analysis with artificial intelligence with few images (from 20)
Create your individual AI-based object or image recognition in up to 30 minutes without programming or expert knowledge – independently and at any time.- Inexpensive – 10x cheaper (based on quotes obtained)
- Performant – Highly accurate with only a few images (from 20)
- Simple – anyone can use it
- Save – Data hosted on own servers behind the Fraunhofer firewall
- For production – Creation of an object recognition without training
DENKweit Offline Analyzer and Operator Software
Integrate our technology into your product or production. Use our modular offline analyzer or our APIs/DLLs
- Modular adaptable to your needs. We optimize our solution to your wishes.
- No connection to the Internet required during operation. Evaluation in “ms” locally on the device.
- Integrate our technology into your solutions via API/DLL
- No data business model – we don’t collect data!
https://denkweit.com/en/denknetze-landing-page/denknetze-technology/
news from 08.03.2023:
Fraunhofer spin-off DENKweit and IDS announce sales collaboration
Making AI easily accessible together
Fraunhofer Spin-Off DENKweit und IDS geben Vertriebskooperation bekannt
Gemeinsam KI einfach zugänglich machen
IDS NXT | Industrial cameras with AI | Complete system | Easy and fast
AI Vision for beginners and professionals: Find out how IDS NXT supports you in image processing with artificial intelligence.
en.ids-imaging.com
This is how Deep Learning can be used quickly, easily and cost-effectively - by everyone in all industries. IDS NXT is the all-in-one system for the use of intelligent cameras. You benefit from user-friendly workflows and perfectly coordinated hardware and software.
They are linked to edge-ai + vision alliance:
2D, 3D and AI: IDS Presents Numerous New Products and Camera Developments at VISION
Today, cameras are often more than just suppliers of images – they can recognize objects, generate results or trigger follow-up processes. Visitors to
www.edge-ai-vision.com
M_C
Founding Member
Tachyum on LinkedIn: #ai #ew23 #edge #embedded #iot #software #ip
Embedded World 2023: VDC’s “Embeddy” Award Winners - Akida 2nd Generation: Direct support for the embedded AI processor market, which is rapidly growing…
Getupthere
Regular
COVID or flu? Ultra-thin sensor 'can detect both within 10 seconds', say scientists — Sky News
Scientists have developed a sensor that can test for COVID and flu at the same time and tell which one you have within 10 seconds.
Fact Finder
Top 20
Great find generously shared:
“Embedded World 2023: VDC’s “Embeddy” Award Winners
by Dan Mandell & Chris Rommel | 3/28/2023Last week, VDC Research attended the Embedded World 2023 conference in Nuremberg, Germany and held dozens of face-to-face meetings with exhibitors. Following a barren 2020 and relatively muted attendance last year, Embedded World 2023 was a clear indication that this industry is a small step away from back to normal conference and tradeshow attendance and activities. The halls were packed and bustling like times of old, with everyone sharing a sense of relief and rejuvenated appreciation for in-person events, meetings, and connections.
VDC created and named the Embeddy Awards to highlight companies announcing important advances in the IoT and embedded software, hardware, and services industries. The Embeddy is awarded for the most cutting-edge product or service available to embedded software developers and system engineers. Nominees were judged across specific criteria including corporate, technological, and industry significance, as well as the most cutting-edge hardware and software solutions or services.
The 2023 Embeddy Award Winners include:
- IP: BrainChip Akida 2nd Generation
- HARDWARE: Eurotech ReliaCOR Secure Edge AI
- SOFTWARE: 1NCE OS
Embeddy IP Winner: BrainChip
Akida 2nd Generation: Leading up to Embedded World 2023, BrainChip, a commercial producer of ultra-low power neuromorphic AI IP, launched the 2nd generation of its Akida platform to drive provide direct support for the embedded AI processor market, which is rapidly growing, particularly at the far edge. More complex networks like RESNET-50 can be processed by Akida without CPU intervention and the platform supports 8-bit weights to go with existing 4,2,1 bit support. Akida also adds an efficient vision transformer implementation for acceleration and can also handle 3D data for applications like video object detection and target tracking as well as 1D time series data for streaming data on devices with limited memory, battery, and compute resources. Akida features separable spatial temporal convolutions called Temporal Event Based Neural Nets that allow for simpler and efficient AI processing solutions.”
Boab
I wish I could paint like Vincent
I believe this is likely the communications company we are involved with.....this is one of many brn/ teal interactions
Imo dyor
tealcom.io
Imo dyor
What Are Common IoT Use Cases That Require Cellular Connectivity?
The Internet of Things (IoT) is transforming the way we […]
tealcom.io
Fact Finder
Top 20
I don’t often let myself expect too much from these promotions but I would actually bet on Brainchip being involved unless Teksun have secretly found another first in class Ai Accelerator and Edge processor since they announced Brainchip AKIDA and not Cisco and Toshiba and I trolled the depths of their website:
AI & Machine Learning Services | AI and ML Services | Product Engineering Company
We have fostered a broad spectrum of AI and ML Services right from the NLP to the knowledge virtualization to Smart Analytics, Predictive Analysis, RPA, and many more.
teksun.com
Then finding:
MegaChips is a semiconductor company that designs and manufactures various integrated circuits and system-on-chip (SoC) products. MegaChips provides various SoC products for applications, including the Internet of Things (IoT), and automotive, industrial, and consumer electronics. Their products include microcontrollers, wireless communication modules, image sensors, audio processors, and power management ICs.
One notable product from MegaChips is the BlueChip Wireless SoC, which provides low-power wireless connectivity for IoT devices using Bluetooth Low Energy (BLE) technology. The BlueChip SoC supports many devices, including sensors, beacons, and wearable devices. MegaChips also offers a range of automotive products, including advanced driver-assistance systems (ADAS) and infotainment systems. Their products are used in various automotive applications, such as parking assistance, collision avoidance, and vehicle-to-vehicle (V2V) communication
Edge Impulseis a platform that enables developers to build and deploy machine learning models on small devices like microcontrollers without requiring deep expertise in machine learning or embedded systems. With Edge Impulse, you can collect data from sensors, preprocess and label it, train a machine learning model, and deploy the model to your device -- all through a simple web interface. The platform supports a variety of sensors and development boards and provides pre-built libraries for many popular devices, making it easy to get started.
Edge Impulse is particularly useful for building applications requiring low latency and high accuracies, such as real-time signal processing, computer vision, and predictive maintenance. It can be used in various industries, including automotive, manufacturing, and healthcare.
The platform is available in both a free and paid version, with the free version allowing up to 1,000 data samples per month and the paid version providing more features and higher data limits
Renesas Electronics Corporation is a Japanese semiconductor manufacturer specializing in providing microcontroller and microprocessor solutions for various industries, including automotive, industrial, home electronics, and information communication technology. Renesas's automotive products are used in various applications, including advanced driver assistance systems (ADAS), powertrain control, and infotainment systems.
Renesas offers a broad portfolio of products and solutions, including microcontrollers, analog and power devices, sensors, and wireless communication modules. The company's products are designed for various applications, from home appliances and consumer electronics to automotive systems and industrial machinery
And of course:
BrainChipis a technology company making AI everywhere easier to deploy and scale. It’s neuromorphic processor called, Akida™, is pure digital and is an event-based technology that is inherently lower power when compared to conventional neural network accelerators. BrainChip IP supports incremental learning and high-speed inference in a wide variety of use cases, making it cost effective and simple to implement. Designed for sustainable AI, Akida™, performs complex functions at the device level, delivering a real-time, instant response.
BrainChip's solutions have applications in various industries, including automotive, smart homes, security and surveillance. The company has partnered with industry leaders such as Arm, Mercedes Benz, and Renesas driving intelligence into next generation devices
Teksun Partner
As a proven global technology partnerships with vendors and solution providers, we let our customer develop the best of breed solutions for end users.
teksun.com
My opinion only DYOR
FF
AKIDA BALLISTA
Learning
Learning to the Top 🕵♂️
Australian inventors developing autonomous ‘Mr Fix-It’ for satellites
We can all agree that satellites are fragile, expensive, and operate in orbit a long way away from the nearest repair shop.
Learning 🏖
Fact Finder
Top 20
Nerds sell AKIDA too. Straight from the Brainchip website this is what Brainchip expects of its Nerds:
So neuromorphic hardware is exciting! However, it is very difficult to develop and deploy current state-of-the-art solutions onto neuromorphic hardware. It’s also extremely limiting to use existing convolutional neural networks (CNNs) based on these platform models. The difficulty primarily stems from the assumption that neuromorphic hardware is analog, and they only run advanced algorithms with SNNs, which are currently in short supply. Therefore, the production models of today —typically accelerated by traditional Deep Learning Accelerators (DLAs) as a safe path to commercialization—are not supported by neuromorphic hardware. But BrainChip Akida TM is changing the game.
Akida is a fully digital, synthesizable and thus process-independent, and silicon-proven neuromorphic technology. It is designed to be scalable and portable across foundries and architected to be embedded into low-power edge devices. It fully supports acceleration for feed-forward CNNs and accelerates other neural networks like Deep Neural Networks (DNNs), Recurrent Neural Networks (RNNs) and more, while providing the efficiency benefits of a neuromorphic design. It removes risks and simplifies.
BrainChip promotes neuromorphic technology with proven silicon, like AKD1000, but also focuses on enabling end users to use the benefits of this technology with little to no knowledge of neuromorphic science. There are three ways to leverage Akida technology (refer to Figure 2) and deploy complex models:
1. Through BrainChip MetaTF™ framework: It is a unique and free robust ML framework that has Python packages for model development and conversion of TensorFlow/Keras models into Akida. It is a framework that is very popular among AI experts and custom developers. Python packages for MetaTF framework are public, and developers can access the framework here: https://doc.brainchipinc.com
2. Through Edge Impulse studio: It is a unique platform that provides end-to-end development and deployment of Machine Learning models on targeted technology with little to no-code AI expertise. Core functions of BrainChip MetaTF framework are embedded into Edge Impulse studio to deploy models onto Akida targeted silicon, such as AKD1000 SoC. To learn more about how to develop using Edge Impulse, click here or visit https://www.edgeimpulse.com
3. Through Solutions Partners of BrainChip such as NVISO: BrainChip has partnered with solutions providers and enabled them to create complex models using MetaTF and build applications for specific and most common AI use cases such as Human Monitoring solutions. This allows for faster time to market solutions using Akida technology. To learn more about BrainChip Solutions Partners, contact us at sales@brainchip.com.
These avenues provide an opportunity to create and develop a functioning model that is suitable for running on Akida technology. The models are converted using MetaTF and are saved into a serialized byte file. These models can be evaluated offline by running simulations using the Software Runtime provided with MetaTF or can be evaluated on AKD1000 mini PCIe development board using Hardware backend as shown in Figure 3a. Once through the evaluation stage, these models can be deployed into production on any target device with Akida technology. Akida Runtime library, which is a low-level library that is OS agnostic, is used to compile the saved model and inference on any target device that has Akida technology. Customers who license Akida technology for their device are able to compile this Akida
Runtime library with any of their application software and host OS, as shown in Figure 3b.
To learn more about how you can harness the power of AI, request a demo or visit BrainChip.com.
Nikunj Kotecha is a Machine Learning Solutions Architect at BrainChip. With many years of experience and a strong programming background, Kotecha brings a passion for AI-driven solutions to the BrainChip team, with a unique eye for data visualization and analysis. He develops neural networks for neuromorphic hardware and event-based processors and optimizes CNN-based networks for conversion to SNN. Nikunj has an MS in Computer Science from Rochester Institute of Technology.
A multiple choice question.
Who is Brainchip selling its AKIDA technology IP too?
1. The person in the street;
2. Homemakers;
3. Shareholders;
4. other technology Nerds building Edge Devices.
My opinion only DYOR
FF
AKIDA BALLISTA
Developing CNNs for Neuromorphic Hardware
(Been There! Done That!)
By Nikunj Kotecha
Often, we hear that neuromorphic technology is cool, classy, low power, next-gen hardware for AI and the most suitable technology for edge devices. Neuromorphic technology mimics the brain, the most efficient computation engine known, to create a computing and natural learning paradigm for devices. Neuromorphic design is complemented with Spiking Neural Networks (SNNs) which emulate how neurons fire and hence only compute when absolutely necessary. This is unlike today’s “MAC monsters" engines -ones that execute lots of MACs (Multiply Accumulate operations which are the basis of most AI computation) in parallel, many of which often get discarded.So neuromorphic hardware is exciting! However, it is very difficult to develop and deploy current state-of-the-art solutions onto neuromorphic hardware. It’s also extremely limiting to use existing convolutional neural networks (CNNs) based on these platform models. The difficulty primarily stems from the assumption that neuromorphic hardware is analog, and they only run advanced algorithms with SNNs, which are currently in short supply. Therefore, the production models of today —typically accelerated by traditional Deep Learning Accelerators (DLAs) as a safe path to commercialization—are not supported by neuromorphic hardware. But BrainChip Akida TM is changing the game.
Akida is a fully digital, synthesizable and thus process-independent, and silicon-proven neuromorphic technology. It is designed to be scalable and portable across foundries and architected to be embedded into low-power edge devices. It fully supports acceleration for feed-forward CNNs and accelerates other neural networks like Deep Neural Networks (DNNs), Recurrent Neural Networks (RNNs) and more, while providing the efficiency benefits of a neuromorphic design. It removes risks and simplifies.
Figure 1. Akida Tech Foundations. Fundamentally different and extremely efficient.
At BrainChip, our mission is to Unlock the future of AI. We believe we can do this by enabling Edge devices with Akida, next-generation technology to advance the growth and smartness of these devices and ultimately provide end users with a sense of privacy and security, energy and cost savings while having access to new features. We realize to achieve our mission, we must make it easy for our end users (who may be experts or non-experts in the field of AI) to use our technology and support current solutions. BrainChip provides development boards of their reference chip AKD1000 for anybody in the community to use and build prototype modules. For commercial use, BrainChip provides licenses of the technology to get it integrated into a custom ASIC, board, or a module that can be used in millions of edge devices.BrainChip promotes neuromorphic technology with proven silicon, like AKD1000, but also focuses on enabling end users to use the benefits of this technology with little to no knowledge of neuromorphic science. There are three ways to leverage Akida technology (refer to Figure 2) and deploy complex models:
Figure 2. BrainChip development ecosystem and access to deployment of solutions to Akida technology
1. Through BrainChip MetaTF™ framework: It is a unique and free robust ML framework that has Python packages for model development and conversion of TensorFlow/Keras models into Akida. It is a framework that is very popular among AI experts and custom developers. Python packages for MetaTF framework are public, and developers can access the framework here: https://doc.brainchipinc.com
2. Through Edge Impulse studio: It is a unique platform that provides end-to-end development and deployment of Machine Learning models on targeted technology with little to no-code AI expertise. Core functions of BrainChip MetaTF framework are embedded into Edge Impulse studio to deploy models onto Akida targeted silicon, such as AKD1000 SoC. To learn more about how to develop using Edge Impulse, click here or visit https://www.edgeimpulse.com
3. Through Solutions Partners of BrainChip such as NVISO: BrainChip has partnered with solutions providers and enabled them to create complex models using MetaTF and build applications for specific and most common AI use cases such as Human Monitoring solutions. This allows for faster time to market solutions using Akida technology. To learn more about BrainChip Solutions Partners, contact us at sales@brainchip.com.
These avenues provide an opportunity to create and develop a functioning model that is suitable for running on Akida technology. The models are converted using MetaTF and are saved into a serialized byte file. These models can be evaluated offline by running simulations using the Software Runtime provided with MetaTF or can be evaluated on AKD1000 mini PCIe development board using Hardware backend as shown in Figure 3a. Once through the evaluation stage, these models can be deployed into production on any target device with Akida technology. Akida Runtime library, which is a low-level library that is OS agnostic, is used to compile the saved model and inference on any target device that has Akida technology. Customers who license Akida technology for their device are able to compile this Akida
Runtime library with any of their application software and host OS, as shown in Figure 3b.
Figure 3a. Using MetaTF Software backend for simulations and Hardware backend for model deployment on AKD1000 mini PCIe development board
Figure 3b. Using low-level Akida Runtime Library for production deployment of models in target devices with Akida technology
BrainChip is very excited about the ecosystem that is available for our end users to use, develop, and deploy complex AI models on Akida neuromorphic technology. Expert AI developers who are familiar with CNN architectures can use BrainChip MetaTF framework to deploy familiar models on Akida. Developers with little to no code experience can use Edge Impulse studio to deploy models on Akida technology and users who want faster time to market can work with Solutions Partners such as NVISO.To learn more about how you can harness the power of AI, request a demo or visit BrainChip.com.
Nikunj Kotecha is a Machine Learning Solutions Architect at BrainChip. With many years of experience and a strong programming background, Kotecha brings a passion for AI-driven solutions to the BrainChip team, with a unique eye for data visualization and analysis. He develops neural networks for neuromorphic hardware and event-based processors and optimizes CNN-based networks for conversion to SNN. Nikunj has an MS in Computer Science from Rochester Institute of Technology.
A multiple choice question.
Who is Brainchip selling its AKIDA technology IP too?
1. The person in the street;
2. Homemakers;
3. Shareholders;
4. other technology Nerds building Edge Devices.
My opinion only DYOR
FF
AKIDA BALLISTA
Good morning,
Great to see that we are receiving more exposure, Akida IP 2nd Gen now being recognized as a real winner among our peers !
My neighbor who attended the embedded conference in Germany said it was a massive event, winning this award will pull more
clients into our IP world, congratulations Peter, Anil and the entire Brainchip family.
I asked a certain someone if they would be attending the AGM in May this year, but not to be, for all the right reasons.
Work ethic says it all
"I am too busy with next generation work on the AKIDA IP, it almost take a full week away that I cant afford at this stage"
Proud to be associated with staff of this quality, who are ultimately working hard for the benefit of all.
Love Brainchip x
Tech
Great to see that we are receiving more exposure, Akida IP 2nd Gen now being recognized as a real winner among our peers !
My neighbor who attended the embedded conference in Germany said it was a massive event, winning this award will pull more
clients into our IP world, congratulations Peter, Anil and the entire Brainchip family.
I asked a certain someone if they would be attending the AGM in May this year, but not to be, for all the right reasons.
Work ethic says it all
"I am too busy with next generation work on the AKIDA IP, it almost take a full week away that I cant afford at this stage"
Proud to be associated with staff of this quality, who are ultimately working hard for the benefit of all.
Love Brainchip x
Tech
Fact Finder
Top 20
Anyone else notice that Amazon Web Services were a nominee for an Embeddy Award 2023:
AWS IoT Core for Amazon Sidewalk
Exhibitor: AWS
Hall/Booth: 4/4-550
Due to the high connectivity cost, power consumption, or limited range and coverage of existing networks, innovation of IoT solutions from developers has historically been limited, which has resulted in narrow availability for end users. While cellular companies have wide network coverage, the higher cost of this coverage reduces the feasibility of many use cases.
Other technologies (e.g. LoRaWAN) that offer low-cost and high-power solutions don’t provide the necessary network coverage or security. Similarly, technologies like WiFi, BLE, Thread, ZigBee, and Z-Wave are well-established smart home solutions, but fall short due to limited range when connectivity is required beyond the home.
Today, IoT devices drop off the internet at astonishingly high rates and frequently never get reconnected, which in turn drives reliability and performance issues for IoT products in the field.
Amazon Sidewalk is a shared network that helps devices like the Amazon Echo, Ring security cameras, and motion sensors work better at home and beyond the front door. When enabled, the network can support other Sidewalk devices in your community, and can be used for applications such as sensing your environment and alerting you when there's a water leak.
Amazon Sidewalk provides redundant coverage for many devices on the network. Therefore, when a Sidewalk device becomes disconnected from one gateway, it can re-establish connectivity by automatically connecting to another available gateway; no intervention is required of end-user. The typical range for many Amazon Sidewalk bridges is one half mile/one kilometer. AWS IoT Core for Amazon Sidewalk provides cloud services that you can use to connect the Sidewalk devices to the AWS Cloud and use other AWS services.
With AWS IoT Core for Amazon Sidewalk, you can build intelligent applications that are capable of increasing the efficiencies across all types of facilities. Sidewalk-enabled sensors can be deployed across buildings, cities, or other types of infrastructure to monitor and control smart systems. With instant connect capabilities, these Sidewalk-enabled devices can simply ‘power-on’ and immediately begin sending data to the cloud. No complex app setup or on-boarding flow required.
Boab
I wish I could paint like Vincent
Brilliant, and good to see they give old mate NVISO a plug. The love in continues.Nerds sell AKIDA too. Straight from the Brainchip website this is what Brainchip expects of its Nerds:
Developing CNNs for Neuromorphic Hardware
(Been There! Done That!)
By Nikunj Kotecha
Often, we hear that neuromorphic technology is cool, classy, low power, next-gen hardware for AI and the most suitable technology for edge devices. Neuromorphic technology mimics the brain, the most efficient computation engine known, to create a computing and natural learning paradigm for devices. Neuromorphic design is complemented with Spiking Neural Networks (SNNs) which emulate how neurons fire and hence only compute when absolutely necessary. This is unlike today’s “MAC monsters" engines -ones that execute lots of MACs (Multiply Accumulate operations which are the basis of most AI computation) in parallel, many of which often get discarded.
So neuromorphic hardware is exciting! However, it is very difficult to develop and deploy current state-of-the-art solutions onto neuromorphic hardware. It’s also extremely limiting to use existing convolutional neural networks (CNNs) based on these platform models. The difficulty primarily stems from the assumption that neuromorphic hardware is analog, and they only run advanced algorithms with SNNs, which are currently in short supply. Therefore, the production models of today —typically accelerated by traditional Deep Learning Accelerators (DLAs) as a safe path to commercialization—are not supported by neuromorphic hardware. But BrainChip Akida TM is changing the game.
Akida is a fully digital, synthesizable and thus process-independent, and silicon-proven neuromorphic technology. It is designed to be scalable and portable across foundries and architected to be embedded into low-power edge devices. It fully supports acceleration for feed-forward CNNs and accelerates other neural networks like Deep Neural Networks (DNNs), Recurrent Neural Networks (RNNs) and more, while providing the efficiency benefits of a neuromorphic design. It removes risks and simplifies.
Figure 1. Akida Tech Foundations. Fundamentally different and extremely efficient.
At BrainChip, our mission is to Unlock the future of AI. We believe we can do this by enabling Edge devices with Akida, next-generation technology to advance the growth and smartness of these devices and ultimately provide end users with a sense of privacy and security, energy and cost savings while having access to new features. We realize to achieve our mission, we must make it easy for our end users (who may be experts or non-experts in the field of AI) to use our technology and support current solutions. BrainChip provides development boards of their reference chip AKD1000 for anybody in the community to use and build prototype modules. For commercial use, BrainChip provides licenses of the technology to get it integrated into a custom ASIC, board, or a module that can be used in millions of edge devices.
BrainChip promotes neuromorphic technology with proven silicon, like AKD1000, but also focuses on enabling end users to use the benefits of this technology with little to no knowledge of neuromorphic science. There are three ways to leverage Akida technology (refer to Figure 2) and deploy complex models:
Figure 2. BrainChip development ecosystem and access to deployment of solutions to Akida technology
1. Through BrainChip MetaTF™ framework: It is a unique and free robust ML framework that has Python packages for model development and conversion of TensorFlow/Keras models into Akida. It is a framework that is very popular among AI experts and custom developers. Python packages for MetaTF framework are public, and developers can access the framework here: https://doc.brainchipinc.com
2. Through Edge Impulse studio: It is a unique platform that provides end-to-end development and deployment of Machine Learning models on targeted technology with little to no-code AI expertise. Core functions of BrainChip MetaTF framework are embedded into Edge Impulse studio to deploy models onto Akida targeted silicon, such as AKD1000 SoC. To learn more about how to develop using Edge Impulse, click here or visit https://www.edgeimpulse.com
3. Through Solutions Partners of BrainChip such as NVISO: BrainChip has partnered with solutions providers and enabled them to create complex models using MetaTF and build applications for specific and most common AI use cases such as Human Monitoring solutions. This allows for faster time to market solutions using Akida technology. To learn more about BrainChip Solutions Partners, contact us at sales@brainchip.com.
These avenues provide an opportunity to create and develop a functioning model that is suitable for running on Akida technology. The models are converted using MetaTF and are saved into a serialized byte file. These models can be evaluated offline by running simulations using the Software Runtime provided with MetaTF or can be evaluated on AKD1000 mini PCIe development board using Hardware backend as shown in Figure 3a. Once through the evaluation stage, these models can be deployed into production on any target device with Akida technology. Akida Runtime library, which is a low-level library that is OS agnostic, is used to compile the saved model and inference on any target device that has Akida technology. Customers who license Akida technology for their device are able to compile this Akida
Runtime library with any of their application software and host OS, as shown in Figure 3b.
Figure 3a. Using MetaTF Software backend for simulations and Hardware backend for model deployment on AKD1000 mini PCIe development board
Figure 3b. Using low-level Akida Runtime Library for production deployment of models in target devices with Akida technology
BrainChip is very excited about the ecosystem that is available for our end users to use, develop, and deploy complex AI models on Akida neuromorphic technology. Expert AI developers who are familiar with CNN architectures can use BrainChip MetaTF framework to deploy familiar models on Akida. Developers with little to no code experience can use Edge Impulse studio to deploy models on Akida technology and users who want faster time to market can work with Solutions Partners such as NVISO.
To learn more about how you can harness the power of AI, request a demo or visit BrainChip.com.
Nikunj Kotecha is a Machine Learning Solutions Architect at BrainChip. With many years of experience and a strong programming background, Kotecha brings a passion for AI-driven solutions to the BrainChip team, with a unique eye for data visualization and analysis. He develops neural networks for neuromorphic hardware and event-based processors and optimizes CNN-based networks for conversion to SNN. Nikunj has an MS in Computer Science from Rochester Institute of Technology.
A multiple choice question.
Who is Brainchip selling its AKIDA technology IP too?
1. The person in the street;
2. Homemakers;
3. Shareholders;
4. other technology Nerds building Edge Devices.
My opinion only DYOR
FF
AKIDA BALLISTA
Stable Genius
Regular
Tachyum on LinkedIn: #ai #ew23 #edge #embedded #iot #software #ip
Embedded World 2023: VDC’s “Embeddy” Award Winners - Akida 2nd Generation: Direct support for the embedded AI processor market, which is rapidly growing…www.linkedin.com
View attachment 33131
There’s a lot going on with Tachyum. They offer SNN for data centres as a licensable IP core but I’m trying to decipher if it’s software or hardware.
Prodigy: The World's First Universal Processor | Tachyum
Tachyum is enabling human brain-scale AI and advancing the entire world to a greener era, by delivering Tachyum Prodigy - The World’s First Universal Processor
Tachyum Prodigy integrates the functionality of CPUs, GPUs and TPUs into a single homogeneous architecture, delivering cutting-edge performance, energy consumption, server utilization, and space efficiency to address the growing demands of AI, HPC, and hyperscale data centers.
Last edited:
Boab
I wish I could paint like Vincent
I see our resident verification engineer is on the job.
Similar threads
