wilzy123
Founding Member
No...... it's more fun to speculate, and then extrapolate that speculation out to even bigger claims and then start hyperventilating
BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
No...... it's more fun to speculate, and then extrapolate that speculation out to even bigger claims and then start hyperventilating
SERA2g
Founding Member
Serious news!
BrainChip Names Former Arm Executive Nandan Nayampally as Chief Marketing Officer
Sunday, November 20, 2022 5:30 PM
Brainchip Holdings Limited/ADR
Home
Unlock the power of AI with BrainChip. Enhance data processing, Edge apps and neural networks at the speed of tomorrow. Explore now!brainchip.com
Share this Article
Topic:
Management Changes
LAGUNA HILLS, CA / ACCESSWIRE / November 20, 2022 /BrainChip Holdings Ltd (ASX:BRN)(OTCQX:BRCHF)(ADR:BCHPY), the world's first commercial producer of neuromorphic AI IP, has hired Nandan Nayampally as Chief Marketing Officer to drive all aspects of marketing, product management and business strategy for the company's AkidaTM event-based AI neural processor IP, and its portfolio of Essential AI enabling technology solutions.
Nayampally is an entrepreneurial executive with more than 25 years of success in building or growing technology businesses with industry-wide impact. He was at Arm for more than 15 years in a variety of product marketing and product management leadership roles, eventually becoming vice president and general manager of Arm's signature CPU group and the Client Line of Business, , where he identified key technology investments, developed a strategy and roadmap of products to deliver compelling, market-leading solutions for billions of SoCs, while establishing strategic partnerships and alliances. Nayampally comes to BrainChip from Amazon, where he helped accelerate the adoption of Alexa Voice and other multimodal services into third-party devices.
"I am excited to join BrainChip. Our unique approach to performant and efficient edge AI at scale is a great enabler for an industry that is looking for innovative and transformative solutions" said Nayampally. "It is a great opportunity to not only advance product intelligence at the sensor and the edge but unleash the full power of AI. BrainChip is positioned to create that positive change and I'm thrilled to be a part of making that happen."
"Nandan has a deep technical understanding of semiconductor design and IP as well as the market factors that lead to product success," said Sean Hehir, CEO. "We look forward to leveraging his product experience and executive success at companies like Arm and Amazon to help us deliver BrainChip solutions to the market."
" and its portfolio of Essential AI enabling technology solutions."
"Portfolio" would generally mean more than one.
We must be getting close to hearing more about 2.0.
IMO DYOR
Rskiff
Regular
and latest from a ICE vehicle. https://www.slashgear.com/1110627/jay-leno-reveals-graphic-details-about-his-car-fire-burns/
JoMo68
Regular
I think Jerome had a pretty serious cycling accident not so long ago (he spoke about it in one of his podcasts - not sure which one though) - maybe taking a bit of a step back?
Fullmoonfever
Top 20
Whilst the below April 2022 independent testing article in Plos One utilises Loihi, like to believe we all have an understanding of how we stack up against Loihi.
The main takeaways for me and the supporting results only add weight to the potential and capabilities of Akida vs CPU & GPU and how it provides an exceptional complimentary accelerator / processor imo.
Full article here for those more tech savvy:
journals.plos.org
The comparison of the average power consumptions between the SNNs and the ANNs is shown in Table 4. With the batch size set to one, the SNN with 16 time steps uses 217.0x/24.0x less power than the ANNs on Xeon/i7 CPUs, 9.3x less than the ANN on ARM CPU, and 40.8x/31.3x less than the ANNs on V100/T4 GPU. This is where neuromorphic hardware starts to shine as it consumes way less power than the conventional hardware. Utilizing the temporal sparsity of SNN appropriately, we believe the neuromorphic hardware can further reduce its power consumption. Another thing that we can observe from Table 4 is that the static (idle) power dominates the power consumption of the Loihi chip.
We measured the total energy used per inference (forward pass) reported in Table 5. These results can also be estimated by combining the results of Tables 3 and 4. As summarized in Table 5, with the batch size set to one, the energy consumption of SNN with 16 time steps is 15.6x/3.2x less than the ANNs on Xeon/i7 CPUs, 2.5x less than the ANN on ARM CPU, and 17.5x/12.5x less than the ANNs on V100/T4 GPUs per inference. This proves the benefits of the neuromorphic hardware in the low energy-budget applications of machine learning, particularly lightweight image search engines and visual recommender systems. It is apparent that when large batch sizes are used, CPUs and GPUs consume less energy per example. However, there are many use cases where inference is executed in small batches, and they are the targets for neuromorphic hardware in the current stage.
The main takeaways for me and the supporting results only add weight to the potential and capabilities of Akida vs CPU & GPU and how it provides an exceptional complimentary accelerator / processor imo.
Full article here for those more tech savvy:
Neuromorphic computing for content-based image retrieval
Neuromorphic computing mimics the neural activity of the brain through emulating spiking neural networks. In numerous machine learning tasks, neuromorphic chips are expected to provide superior solutions in terms of cost and power efficiency. Here, we explore the application of Loihi, a...
Abstract
Neuromorphic computing mimics the neural activity of the brain through emulating spiking neural networks. In numerous machine learning tasks, neuromorphic chips are expected to provide superior solutions in terms of cost and power efficiency. Here, we explore the application of Loihi, a neuromorphic computing chip developed by Intel, for the computer vision task of image retrieval. We evaluated the functionalities and the performance metrics that are critical in content-based visual search and recommender systems using deep-learning embeddings. Our results show that the neuromorphic solution is about 2.5 times more energy-efficient compared with an ARM Cortex-A72 CPU and 12.5 times more energy-efficient compared with NVIDIA T4 GPU for inference by a lightweight convolutional neural network when batch size is 1 while maintaining the same level of matching accuracy. The study validates the potential of neuromorphic computing in low-power image retrieval, as a complementary paradigm to the existing von Neumann architectures.The comparison of the average power consumptions between the SNNs and the ANNs is shown in Table 4. With the batch size set to one, the SNN with 16 time steps uses 217.0x/24.0x less power than the ANNs on Xeon/i7 CPUs, 9.3x less than the ANN on ARM CPU, and 40.8x/31.3x less than the ANNs on V100/T4 GPU. This is where neuromorphic hardware starts to shine as it consumes way less power than the conventional hardware. Utilizing the temporal sparsity of SNN appropriately, we believe the neuromorphic hardware can further reduce its power consumption. Another thing that we can observe from Table 4 is that the static (idle) power dominates the power consumption of the Loihi chip.
We measured the total energy used per inference (forward pass) reported in Table 5. These results can also be estimated by combining the results of Tables 3 and 4. As summarized in Table 5, with the batch size set to one, the energy consumption of SNN with 16 time steps is 15.6x/3.2x less than the ANNs on Xeon/i7 CPUs, 2.5x less than the ANN on ARM CPU, and 17.5x/12.5x less than the ANNs on V100/T4 GPUs per inference. This proves the benefits of the neuromorphic hardware in the low energy-budget applications of machine learning, particularly lightweight image search engines and visual recommender systems. It is apparent that when large batch sizes are used, CPUs and GPUs consume less energy per example. However, there are many use cases where inference is executed in small batches, and they are the targets for neuromorphic hardware in the current stage.
Conclusion
We studied the application of the Loihi chip, a neuromorphic computing hardware developed by Intel, in image retrieval. Our results show that the generation of the deep learning embeddings by spiking neural networks for lightweight convolutional neural networks is about 2.5 times more energy-efficient compared with a CPU and 12.5 times more energy-efficient compared with a GPU. We confirm the long-term potential of neuromorphic computing in machine learning, not as a replacement for the predominant von Neumann architecture, but as accelerated coprocessors.Fact Finder
Top 20
Yes he had two.
The first one was quite serious with multiple broken ribs and yet he continued to carry out his role for Brainchip.
He then had a second one breaking more ribs. Cycling is apparently his first love and reason for being and was always going to be his destination.
Those who have never found their one thing will never understand why those who have do what they do. Not everything is about money.
My opinion only DYOR
FF
AKIDA BALLISTA
I think Nandan coming on board is a brilliant appointment - from Amazon no less. That in itself is impressive as one would expect a former employee of Jeff Bezos wouldn't be doing it for the money.
What is even more telling is that to be offered and subsequently accept an executive position at Brainchip one would have been shown what's happening under the hood. The people we are attracting seem to be the type that are able to recognise the significance of what they are seeing and despite their experience and exposure to industry technology they continue to be impressed.
What is even more telling is that to be offered and subsequently accept an executive position at Brainchip one would have been shown what's happening under the hood. The people we are attracting seem to be the type that are able to recognise the significance of what they are seeing and despite their experience and exposure to industry technology they continue to be impressed.
Fact Finder
Top 20
Others may be able to confirm this but second hand I am able to suggest that a single Loihi 2 chip requires 2 to 3 watts to do its thing.Whilst the below April 2022 independent testing article in Plos One utilises Loihi, like to believe we all have an understanding of how we stack up against Loihi.
The main takeaways for me and the supporting results only add weight to the potential and capabilities of Akida vs CPU & GPU and how it provides an exceptional complimentary accelerator / processor imo.
Full article here for those more tech savvy:
Neuromorphic computing for content-based image retrieval
Neuromorphic computing mimics the neural activity of the brain through emulating spiking neural networks. In numerous machine learning tasks, neuromorphic chips are expected to provide superior solutions in terms of cost and power efficiency. Here, we explore the application of Loihi, a...
Abstract
Neuromorphic computing mimics the neural activity of the brain through emulating spiking neural networks. In numerous machine learning tasks, neuromorphic chips are expected to provide superior solutions in terms of cost and power efficiency. Here, we explore the application of Loihi, a neuromorphic computing chip developed by Intel, for the computer vision task of image retrieval. We evaluated the functionalities and the performance metrics that are critical in content-based visual search and recommender systems using deep-learning embeddings. Our results show that the neuromorphic solution is about 2.5 times more energy-efficient compared with an ARM Cortex-A72 CPU and 12.5 times more energy-efficient compared with NVIDIA T4 GPU for inference by a lightweight convolutional neural network when batch size is 1 while maintaining the same level of matching accuracy. The study validates the potential of neuromorphic computing in low-power image retrieval, as a complementary paradigm to the existing von Neumann architectures.
The comparison of the average power consumptions between the SNNs and the ANNs is shown in Table 4. With the batch size set to one, the SNN with 16 time steps uses 217.0x/24.0x less power than the ANNs on Xeon/i7 CPUs, 9.3x less than the ANN on ARM CPU, and 40.8x/31.3x less than the ANNs on V100/T4 GPU. This is where neuromorphic hardware starts to shine as it consumes way less power than the conventional hardware. Utilizing the temporal sparsity of SNN appropriately, we believe the neuromorphic hardware can further reduce its power consumption. Another thing that we can observe from Table 4 is that the static (idle) power dominates the power consumption of the Loihi chip.
View attachment 22541
We measured the total energy used per inference (forward pass) reported in Table 5. These results can also be estimated by combining the results of Tables 3 and 4. As summarized in Table 5, with the batch size set to one, the energy consumption of SNN with 16 time steps is 15.6x/3.2x less than the ANNs on Xeon/i7 CPUs, 2.5x less than the ANN on ARM CPU, and 17.5x/12.5x less than the ANNs on V100/T4 GPUs per inference. This proves the benefits of the neuromorphic hardware in the low energy-budget applications of machine learning, particularly lightweight image search engines and visual recommender systems. It is apparent that when large batch sizes are used, CPUs and GPUs consume less energy per example. However, there are many use cases where inference is executed in small batches, and they are the targets for neuromorphic hardware in the current stage.
View attachment 22542
Conclusion
We studied the application of the Loihi chip, a neuromorphic computing hardware developed by Intel, in image retrieval. Our results show that the generation of the deep learning embeddings by spiking neural networks for lightweight convolutional neural networks is about 2.5 times more energy-efficient compared with a CPU and 12.5 times more energy-efficient compared with a GPU. We confirm the long-term potential of neuromorphic computing in machine learning, not as a replacement for the predominant von Neumann architecture, but as accelerated coprocessors.
Which if confirmed as correct would make AKIDA’s power consumption so tiny by comparison to be hardly a factor.
Reminds years and years ago now 4 Corners or it’s predecessor did a story about Australian industry moving manufacturing off shore to China.
They interviewed the owner of the Bissell Brush Company and he said that it was labour cost that caused their move. He stated that in Australia labour accounted for 40% of the cost of a single brush whereas at their factory in China it was not able to be costed as a percentage of an individual brush as it was a fraction of one cent.
Why I recite this story is simply if the power savings are great enough the market will come and there is no doubt Brainchip is to the semiconductor industry what China was back then to manufacturing.
My opinion only DYOR
FF
AKIDA BALLISTA
Shadow59
Regular
FF, whilst I'm a great fan of all you write, unfortunately this time however you are writing about something you do not understand.
If Jerome had fallen and broken some ribs riding his bike (on a few occaisions), I'm pretty sure it would have no affect on how and when he worked for Brainchip and how he did his job. The reason if anything would be so he had more time to train for competition( though he could potentially break even more ribs and do more damage, thats what pain killers are for.) He is most likely in it for the adrenalin rush which is like an addiction. It hypes the mind makes you positive and improves the quality of your work. But it does not make you want to retire from life and what you do.
"Not everything is about money". Say what????
Wrong, ....FF !!!
If you do not have money, then EVERYTHING is about money. Do you not remember what it was like to not have money? If you have always had money or the means, you cannot understand that people who do not have the resources could care less about finding their one thing, or what I call one's bliss. You were a policeman at one time, right? You should know that most of the world cannot even begin to dream of finding "their one thing", especially if it involves the purchase of a high end bike, .....which if Mr. Nadel does not own I will eat my words.
Food is billions of people's "one thing".
To say not everything is about money is quite elitist of you, frankly. I am surprised, ..... you are not that out of touch, imo.
However, that is your opinion and you are entitled to it. I just think it's disingenuous. Let's all guess how much Mr. Nadel's bicycle cost? I'll start,.....my guess is $5,000
If it's not about money, then he would enjoy riding a rusted 50 year old Schwinn one speed as much as he enjoys his current ride.....and that I can assure you is very unlikely, and that is my opinion. Since we are both simply stating opinions I will go now. having voiced mine.
I hope I have not poked a hornets nest. Regards, dippY
I think I may be getting somewhere with Cambridge Consultants. Part of the ARM ecosystem and design partner. Some very interesting links here but it’s starting to get above my pay grade and need some of the more technically minded here to have a quick look.
www.arm.com
www.arm.com
The applications for this research are huge: real-time medical AI can be deployed in low power endpoint devices and used in settings where Internet connectivity is unavailable, or bulky and power-hungry computing equipment would be impractical.
It also opens the door to combine further benefits of processing AI data on endpoints, including lower latency and lower power. All whilst leaving room to improve privacy and security, since data does not leave the users device.
Cambridge Consultants
Cambridge Consultants creates and delivers breakthrough products and services for clients worldwide.
www.arm.com
Cambridge Consultants: How We're Pushing the Endpoint AI Envelope
Cambridge Consultants shares how the company achieved an incredible 7x power reduction, 1,000x speed increase in Arm-powered endpoint AI
www.arm.com
The applications for this research are huge: real-time medical AI can be deployed in low power endpoint devices and used in settings where Internet connectivity is unavailable, or bulky and power-hungry computing equipment would be impractical.
It also opens the door to combine further benefits of processing AI data on endpoints, including lower latency and lower power. All whilst leaving room to improve privacy and security, since data does not leave the users device.
alwaysgreen
Top 20
equanimous
Norse clairvoyant shapeshifter goddess
I wander if he will be announcing the release of the new Valeo bike?
Fullmoonfever
Top 20
Found these couple of basic public projects while sniffing around Edge Studio and the Eon Tuner which was mentioned on those earlier videos I posted.
If anyone likes to do some other digging around Edge Impulse.
studio.edgeimpulse.com
Notice the target on this one
Here's another.
Notice the target on this one
Belongs to these guys.....obviously doesn't mean we are working with them but does infer that our NN works on that device I would suggest
Some of the device choices in the drop down......who knows what our NN can work with
If anyone likes to do some other digging around Edge Impulse.
Brainchip / Akida Image Classification
This is the finished Edge Impulse project for the tutorial 'Adding sight to your sensors'. From here you can acquire new training data, design impulses and train models.
studio.edgeimpulse.com
Notice the target on this one
Here's another.
Notice the target on this one
Belongs to these guys.....obviously doesn't mean we are working with them but does infer that our NN works on that device I would suggest
Some of the device choices in the drop down......who knows what our NN can work with
Fact Finder
Top 20
Herein lies the problem you simply do not have sufficient facts available to you unless you have insider knowledge to say as an immutable FACT that “There is simply no excuse”.
You might have said you cannot think of an excuse but in my eyes having chosen the words you have you loose all credibility even though I could give weight to some of your other observations.
Yet even there the statement “keeps shareholders fully informed.”
What does he release details of all customer engagements and the precise commercial terms of agreements? I would think not.
Does he make compulsory price sensitive announcements required by the ASX well so does Brainchip.
So what you are probably saying is he releases enough information that you are kept happy.
I too can say Brainchip releases enough information that I am kept happy.
There is little value in these two statements however because happy is very subjective.
So can you explain why you being unhappy is in and of itself of sufficient importance that Brainchip’s approach is an issue of poor management?
There are 46,000 shareholders you and I are but two.
If you are unhappy about the present share price I too am unhappy about the present share price but I would need to be convinced that the present share price is a product of not releasing ‘non price sensitive’ announcements on the ASX that might make some of the 46,000 shareholders happy.
However what is it about a non price sensitive announcement that would make them happy.
Is it the information in the non price sensitive announcement well you and I already know the information from the press release. Yet you are still not happy so it is not the information itself.
So there must be something innate in a ‘non price sensitive’ announcement that elevates the information above its ‘non price sensitive’ import which to me sounds like trying to ramp a share and create hype and false short term value.
My opinion only DYOR
FF
AKIDA BALLISTA
Fact Finder
Top 20
You have not poked a hornets nest rather you have done what you usually do that is take a comment and twist it into something it is not.
The guy was in semi retirement by choice. He clearly had what he considered enough money.
He loves to ride his expensive bike.
He did Sean Hehir a favour coming out of retirement as a favour and having done what was asked he goes back to what he loves doing. If it was all about the money he would still be working at Brainchip. Getting even more money.
I had a very good friend who lived bike riding. He had a bad accident. After he recovered he never went back to bike riding. Why not because he did not place it high enough in his hierarchy of needs to continue.
This guy despite two reasonably serious accidents could not wait to get back on his bike even with the continuing risk of further accidents.
So as I said it’s not all about the money.
But with you it is clearly all about the money.
Professor Fred Hollows when he was alive used to spend the majority of his time doing free eye surgery.
He could have made bundles of money staying here operating on the eyes of people with money.
For him it was not all about the money.
You take every opportunity to be negative and always race to a negative judgment so I can accept that for someone like you it is always about the money.
My opinion only DYOR
FF
AKIDA BALLISTA
Might be the reason why Elon Musk is now moving from battery to Hydrogen fueled vehicles.
Rskiff
Regular
Are you serious ? Hydrogen isn't on his radar for vehicles.The only time he is using hydrogen is to blow up Twitter
Fact Finder
Top 20
I have read my post and your post multiple times and I think you need to read my post again.
You have attributed to my post a meaning that does not exist even in fantasy.
I was confirming the previous posters recollection that he had injured himself in not one but two accidents.
I did not say because of the accidents he could not do his job.
Far from it as I am aware that he continued to do his job despite the accidents and there was some concern expressed to him by some that perhaps he should slow down till he recovered fully.
I know everyone’s level of comprehension is different but seriously read my post again.
My opinion only DYOR
FF
AKIDA BALLISTA
Best not to make assumptions about anyone's position. I got in happily with a 9 cent average thanks.... and have always been confident of BRN's technological lead. First to have commercial silicon etc etc. However, now with other companies i.e. Qualcomm securing contracts with phone manufacturers, Valeo etc using their own SNN technology... it's got me thinking about/questioning BRN's lead? I'm still hopeful. But in the absence of news... it's pretty natural to ask questions. Well at least, that's what I would have thought is reasonable? Instead.. all I get is my comments being dissected, taken out of context and just general abuse from this precious community - who are all satisfied to have blind faith in a highly competative industry.
Similar threads
