thelittleshort
Top Bloke
I mean if BrainChip/USA will not work with China then our IP won’t be in the EQXX in the first place?
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
I mean if BrainChip/USA will not work with China then our IP won’t be in the EQXX in the first place?
Mt09
Regular
Anything could potentially be sold in China at some point, do we not allow our IP to be used in anything? Mercedes are a European company, as long as we’re not selling directly to a CCP controlled company and it’s for non defence related industries we should be right imo.
thelittleshort
Top Bloke
I literally have no idea, I’m just hypothesising
Mt09
Regular
Me either, Sunday morning musings
stockduck
Regular
sounds a little bit exciting for me......
www.eenewseurope.com
"....
The silicon is the NDP120 Neural Decision Processor from Syntiant in California which uses at memory compute and natively processing neural network instructions for enhanced data optimization. This enables InferSens to achieve on-device deep learning processing at only 1% of the power required by traditional processors.
The NDP120 can run multiple AI algorithms simultaneously at under 1mW, among other sensor and voice applications, and is designed to natively run multiple deep neural networks on a variety of architectures, such as CNNs, RNNs and fully connected networks.
The first version of the InferSens’ sensor technology is planned for Q1 2023.
......"
www.syntiant.com
"......
CTO Jonathan Peace will be presenting the technology and the first application at the tinyML EMEA Innovation Forum 2022 in Cyprus on 11th October 2022.
....."
What do you think?
Low power edge AI sensor for smart buildings
Cambridge startup InferSens has teamed up to combine its low power sensor with sophisticated deep learning AI for smart buildings.
www.eenewseurope.com
"....
The silicon is the NDP120 Neural Decision Processor from Syntiant in California which uses at memory compute and natively processing neural network instructions for enhanced data optimization. This enables InferSens to achieve on-device deep learning processing at only 1% of the power required by traditional processors.
The NDP120 can run multiple AI algorithms simultaneously at under 1mW, among other sensor and voice applications, and is designed to natively run multiple deep neural networks on a variety of architectures, such as CNNs, RNNs and fully connected networks.
The first version of the InferSens’ sensor technology is planned for Q1 2023.
......"
InferSens to Introduce Ultra-Low-Power Edge DL Sensor Solution with Syntiant Technology — Syntiant
Battery-Powered Sensor with Embedded Syntiant NDP120 Neural Decision Processor to Enable Highly Accurate On-Device Processing for Commercial and Industrial Applications InferSens CTO Jonathan Peace to Present First Application for Legionella Water System Monitoring at the tinyML EMEA Innovatio
www.syntiant.com
CTO Jonathan Peace will be presenting the technology and the first application at the tinyML EMEA Innovation Forum 2022 in Cyprus on 11th October 2022.
....."
News Release
feeds.issuerdirect.com
What do you think?
Last edited:
equanimous
Norse clairvoyant shapeshifter goddess
All of those executives and engineers can easily migrate to working in India which has better potentials at the moment
Fact Finder
Top 20
Interesting article from April by Sally but does not include Brainchip.
Regards
FF
AKIDA BALLISTA
Fact Finder
Top 20
The USA is critically wounding the Chinese semiconductor industry.
While China is desperately attempting to build its own foundries their plan has a fatal flaw as all of the expertise, equipment and raw materials necessary to design build and maintain the necessary foundries are controlled by the USA either directly or via treaties and political alliances.
So if China wants to put an AKIDA chip in a missile they can buy Mercedes Benz EQXX strip out the AKIDA chip and then what?
This AKIDA chip will not be hardened for extreme conditions necessary to use in missiles etc;
Reverse engineer, design a hardened version then send it to a foundry to be produced? Where?
If Mercedes Benz uses AKIDA in a model and they want to sell that model into China that will be possible.
China will however not be able to ring up Brainchip and say we want ten million hardened AKD1000 chips.
I was reading an article in a defence magazine which claimed the USA in 2021 spent over $800 billion on defence while China spent under $300 billion. When I added in the UK, France, Germany etc the USA spend was over one trillion dollars in 2021.
NATO has 6,000 serviceable nuclear war head missiles in and about Europe and the UK pointed at Russia which claims to have 6,000 pointed in their direction many of which have according to the article not been maintained and are unlikely to launch.
The end outcome of a world war would not lead to China or Russia being victorious. There would be no winners.
I said it when China started to engage in Tiger Diplomacy they had made their run far to early.
China needs the world more than the world needs China. Pingo and Puto have aimed at the West and managed to shoot themselves in the foot.
The US is still the elephant in the room and will be so for the foreseeable future.
My opinion only DYOR
FF
AKIDA BALLISTA
While China is desperately attempting to build its own foundries their plan has a fatal flaw as all of the expertise, equipment and raw materials necessary to design build and maintain the necessary foundries are controlled by the USA either directly or via treaties and political alliances.
So if China wants to put an AKIDA chip in a missile they can buy Mercedes Benz EQXX strip out the AKIDA chip and then what?
This AKIDA chip will not be hardened for extreme conditions necessary to use in missiles etc;
Reverse engineer, design a hardened version then send it to a foundry to be produced? Where?
If Mercedes Benz uses AKIDA in a model and they want to sell that model into China that will be possible.
China will however not be able to ring up Brainchip and say we want ten million hardened AKD1000 chips.
I was reading an article in a defence magazine which claimed the USA in 2021 spent over $800 billion on defence while China spent under $300 billion. When I added in the UK, France, Germany etc the USA spend was over one trillion dollars in 2021.
NATO has 6,000 serviceable nuclear war head missiles in and about Europe and the UK pointed at Russia which claims to have 6,000 pointed in their direction many of which have according to the article not been maintained and are unlikely to launch.
The end outcome of a world war would not lead to China or Russia being victorious. There would be no winners.
I said it when China started to engage in Tiger Diplomacy they had made their run far to early.
China needs the world more than the world needs China. Pingo and Puto have aimed at the West and managed to shoot themselves in the foot.
The US is still the elephant in the room and will be so for the foreseeable future.
My opinion only DYOR
FF
AKIDA BALLISTA
Knowing nothing about chip manufacturing, is it possible to hide functions within a chip which would render it useless if certain predetermined situations arose either autonomously or manually over the cloud?
Fact Finder
Top 20
I don’t know anything either however if I believe what I have read Huawei is alleged to do so all the time.
What I do know is that Peter van der Made has stated that the encryption they have used in AKIDA makes it impossible to reverse engineer the secret sauce.
So at the very least it will not be easy allowing that all things are possible and that Peter van der Made would sensibly talk up the effort they have made to prevent IP theft occurring.
My opinion only DYOR
FF
AKIDA BALLISTA
butcherano
Regular
G’day @Fact Finder. You’ve probably seen this, but Brainchip published the full comparison based on Tiny ML benchmarking at the IESA AI Summit this year which shows that we perform better than Syntiant. This was a month after Sally’s article came out.
From the 6hr 33m 30s mark here….(whole session from Brainchip worth watching though).
Fact Finder
Top 20
Hi @Butcheroo
No I did not see this. I knew that Brainchip had been working on benchmarking early in the year but had not followed up so very grateful to you for pointing this out.
If it is on a slide can you or other more adept hands post a screen shot or copy here.
Again many thanks.
FF
AKIDA BALLISTA
butcherano
Regular
buena suerte :-)
BOB Bank of Brainchip
Might have been posted already ... But a nice write up from our partner ....
Edge Impulse Releases Deployment Support for BrainChip Akida Neuromorphic IP
NEWS PROVIDED BY
Edge Impulse
Aug 03, 2022, 08:00 ET
The tech firms' collaboration augments brain-mimicking Spiking Neural Networks.
SAN JOSE, Calif., Aug. 3, 2022 /PRNewswire/ -- Edge Impulse, the leading platform for enabling ML at the edge, and BrainChip, the leading provider of neuromorphic AI IP technology, announced support for deploying Edge Impulse projects on the BrainChip MetaTF platform.
Edge Impulse enables developers to rapidly build enterprise-grade ML algorithms, trained on real sensor data, in a low to no code environment. These trained algorithms can now be quantized, optimized and converted to Spiking Neural Networks (SNN), which are compatible and can be deployed with BrainChip Akida™ devices. This capability is available for new and existing Edge Impulse projects by using the BrainChip MetaTF model deployment block integrated on the platform. This deployment block enables free-tier developers and enterprise developer users to create and validate neuromorphic models for real-world use-cases and deploy on BrainChip Akida development kits.
A reference project targeting image classification is publicly available as an Edge Impulse public project, giving experts a head start with developing algorithms for the next generation of neuromorphic computing. Machine Learning experts will be able to create Akida-compatible models using the Edge Impulse expert mode within the learning block and using the BrainChip MetaTF model deployment block to use Tensorflow based models, and quantize them to complete or mixed low-precision bits, from 1 to 4 bits. It also allows the implementation of quantization-aware training to help retain the performance. Finally, the optimized model can be converted to Akida-based SNN networks and experts will be able to download them to deploy on BrainChip Akida development kits. Experts also see performance metrics, Akida model summary, and configuration data.
"We're extremely delighted to support BrainChip's proliferation of their Akida technology," said Jan Jongboom, CTO at Edge Impulse. "The combination of BrainChip's advanced technology and Edge Impulse's industry-leading developer experience gives users an easy and seamless way for deploying ML at the edge using efficient and essential SNNs algorithms that helps the advancement of solving real-world problems of all types."
Edge Impulse is widely used to rapidly develop, deploy, and maintain ML algorithms across the industry where security, low-power, and remote deployment needs require intelligence at the edge. BrainChip's Akida IP promises to massively increase the processing capability and efficiency of such applications. Edge Impulse continues to improve access to the BrainChip Akida technology and has started to integrate BrainChip's MetaTF Software Development Environment so that those new to ML and Akida can more easily incorporate the benefits of the technology in their projects.
"Being the commercial leader in edge AI on-chip learning and neuromorphic AI IP availability, it is critical for BrainChip to establish our ecosystem with key partnerships and technology leaders such as Edge Impulse," says Anil Mankar, Chief Development Officer at BrainChip. "Working together, both companies are creating an advantage for the ML community. This is only the beginning of how we see our Akida technology expanding into the ecosystem."
About Edge Impulse
Edge Impulse is the leading machine learning platform, enabling all enterprises to build smarter edge products. Its technology empowers developers to bring more ML products to market faster and helps enterprise teams rapidly develop industry-specific solutions in weeks instead of years. The Edge Impulse platform provides powerful automation and low-code capabilities to make it easier to build valuable datasets and develop advanced ML with streaming data. With over 40,000 developers, and partnerships with the top silicon vendors, Edge Impulse offers a seamless integration experience to validate and deploy with confidence across the largest hardware ecosystem. To learn more, visit edgeimpulse.com.
About BrainChip Inc
BrainChip is the worldwide leader in edge AI on-chip processing and learning. The company's first-to-market neuromorphic processor, AkidaTM, mimics the human brain to analyze only essential sensor inputs at the point of acquisition, processing data with unparalleled efficiency, precision, and economy of energy. Keeping machine learning local to the chip, independent of the cloud, also dramatically reduces latency while improving privacy and data security. In enabling effective edge compute to be universally deployable across real world applications such as connected cars, consumer electronics, and industrial IoT, BrainChip is proving that on-chip AI, close to the sensor, is the future, for its customers' products, as well as the planet. Explore the benefits of Essential AI at brainchip.com.
SOURCE Edge Impulse
Fact Finder
Top 20
When I look at the comparison between AKIDA and Syntiant I am reminded of the CEO Sean Hehir’s warning at the AGM about claims being made by others in the semiconductor space and how they just do not stack up against Brainchips technology.
It is obvious why he stated that AKIDA was the best kept secret in the world and why he is on a mission to educate the world as to the AKIDA revolution.
It is not if but when or Pantene.
Then again is being twice as fast while using 111 times less power really significant???
My opinion only DYOR
FF
AKIDA BALLISTA
butcherano
Regular
Spot on @Fact Finder….and I guess one other benchmark to look at how Brainchip compares with Syntiant would be to look at where Chris Stevens has chosen to work!….When I look at the comparison between AKIDA and Syntiant I am reminded of the CEO Sean Hehir’s warning at the AGM about claims being made by others in the semiconductor space and how they just do not stack up against Brainchips technology.
It is obvious why he stated that AKIDA was the best kept secret in the world and why he is on a mission to educate the world as to the AKIDA revolution.
It is not if but when or Pantene.
Then again is being twice as fast while using 111 times less power really significant???
My opinion only DYOR
FF
AKIDA BALLISTA
Slymeat
Move on, nothing to see.
Not wanting to dampen the brilliant message of Akida being more efficient (as it is) but it gets up my nose to see wording such as "111x less power". More so because of the "111x less" than the fact that it is actually comparing Energy and not Power.
From a purely mathematical point of view 111 x less power is 1 - 111 = -110 meaning it generates 110 times the amount of power. How fantastic is that—when spotting keywords it actually generates power.
I assume they mean it uses 1/111th of the power or 0.9% of the power, which is 99.1% less and not 111x.
Or maybe even it is 99% more efficient!
I won't get into the inaccuracy of calling it a reduction in power though, Joules measure energy, power is measured in Watts or Joules per second. When converting to instantaneous power, the comparison is very different. But it is overall energy used that IS important anyway. And in this chart Akida uses less energy at 5MHz, and more energy at 300Mhz.
So that it is 99% more efficient is correct.
The bar charts appear to give a much better representation—but these are equally confusing in that they are NOT comparing like against like.
But what would be a better thing to report is total bang for buck - as in number of inferences per Watt. i.e. standard scientific normalization to a constant denominator (the Watt).
Let's round for ease of visualizing
ARM Cortex M4 (120MHz) = 150 inf/W
Jetson Nano - (921 MHz) = 1080 inf/W
Jetson Nano (5W - 614 MHz) = 1520 inf/W
Coral Dev Board = 2850 inf/W
AKD 1000 (300MHz) = 17000 inf/W
AKD 1000 (5MHz) = 27000 inf/W
Now we can say roughly 10x the number of inferences for the same power usage, comparing best against the best. And the comparison becomes 180x the number of inferences for the same power usage when comparing best Akida to worst (ARM).
And Also the ARM Cortex looks far inferior now. And I don't think that was the intention of this presentation.
This leads to yet another of my pet hates. Statistics, damned lies, and statistics. "Facts" can be presented any way to show the message you want to convey.
Charts SHOULD ALWAYS be normalized so as to show same versus same.
Similar threads
