BRN Discussion Ongoing

[IloveLamp]

Very possible imo, BRN staff have liked Untether posts in the past.

"There are going to be "tens of thousands of satellites orbiting Earth and [we] can't possibly count on human intervention for them to avoid each other, to pair up with each other or to de-orbit," says Bill Weber, CEO of Firefly Aerospace.

Putting AI devices on satellites could give them more autonomy to navigate using data they collect in real time.
On-board AI processing can also help researchers better leverage the streams of data satellites collect to take the pulse of Earth's forests and fields, monitor methane emissions, and track illegal fishing and other activities and events.

"Sensor data collection is growing exponentially, not only on Earth, but in space, whereas communication downlink technology is only growing linearly," says Paul Quintana of Untether AI. "You can't send all the data from space down anymore. You have to do on-orbit processing."

The bottom line: "We have launch vehicles, launch towers, spacecraft, onboard propulsion and antennas, and in the end, it comes down to the damn semiconductor," Goldin says.

https://www.axios.com/2024/04/26/ai-chips-space-satellites-computing?utm_campaign=editorial&utm_medium=social&utm_source=linkedin

 

[wilzy123]

Gacy has really lost the plot now

 

[Frangipani]

From my post about the upcoming 2024 Telluride Workshop:

Or was it possibly Jens Egholm Pedersen, the Danish doctoral student at Stockholm’s KTH Royal Institute of Technology, Sweden’s largest technical university, who hosted said presentation by Cristian Axenie on the Open Neuromorphic YouTube channel and appeared to be genuinely impressed about Akida (and the Edge Impulse platform), too?

Jens Egholm Pedersen was definitely impressed enough to include Akida as an example of an energy-efficient neuromorphic chip in his latest presentation on neuromorphic computing (stating that it would cost 1 Danish krone (~ 0.22 AUD) per year running it at 100 mW, if I interpret the Danish description alongside the image of the PCIe board correctly - at the same time, he claims Loihi at 1 W would equal 1 DKK per month and Innatera’s ultra-low-power neural processor T1 running at 1 mW would only incur costs of 1 DKK per century):

He also just announced his move from Stockholm to Silicon Valley…

… where he will be doing research under the tutelage of this gentleman:

 

[Damo4]

Where is
View attachment 61646

Him and DH can only appear when the SP is sub 25 cents.
They've left the braincell in good hands for now (Iseki and Dk)

Shame about the trolling. I would rather this forum be silent for a few hours, between interesting titbits, than fill it with pretend "LTH discussion"

 

[Diogenese]

Hi Bravo,

this is what Mohsen Kaboli, Assistant Professor for Robotics and Tactile Intelligence as well as Lead and Principal Investigator of Embodied AI, Robotics, Cognitive Vehicle & Tactile Intelligence at BMW Group commented on Mercedes Chief Software Officer Magnus Östberg’s recent LinkedIn post about neuromorphic computing:

View attachment 61544

View attachment 61546
View attachment 61547

View attachment 61548

View attachment 61549 View attachment 61550

---

In February 2023, you had already come across PHASTRAC and asked Diogenese for his opinion:

---

Here is a recent article on the BMW project Mohsen Kaboli is currently working on: “He is developing a cognitive seat that could appear in BMW vehicles in just a few years.”

https://www.bmw.com/de/freude/wie-sich-die-zukunft-bei-bmw-anfuehlt.html (Deutsch)

https://www.bmw.com/en/freude/a-touch-of-the-future-with-bmw.html (English)

View attachment 61553
View attachment 61554
View attachment 61556

For the sake of completeness, usually one of Frangipanni's fortes, my post in full added:

"Hi Bravo,

Apologies for that German chap.

That project looks to me to be more in the "R" phase of R&D.

"The project aims to develop new devices for implementing the ONN architecture and processing the analog sensor data. It is led by researchers from Technical University in Eindhoven.

It is a university project and universities seem to br enthralled by the close analogy between wetware and analog neurons. They still have to come to grips with the inherent variability of ReRAM/MemRistors.

Not commercial whthin 5 years +. "

Just to expand on that,

https://phastrac.eu/

PHASTRAC is a research project funded by Horizon EU’s research and innovation programme with core subject “Phase Transition Materials for Energy Efficient Edge Computing”. The project with duration of 42 months (1 January 2023 – 30 June 2026) brings together leading European research and academic institutions.

PHASTRAC aims to develop a novel analog-to-information neuromorphic computing paradigm based on oscillatory neural networks (ONNs). We offer a first-of-its-kind and novel analog ONN computing architecture to seamlessly interface with sensors and process their analog data without any analog-to-digital conversion. ONNs are a biologically inspired neuromorphic computing architecture, where neuron oscillatory behavior will be developed by innovative phase change VO2 material coupled with synapses developed by bilayer Mo/HfO2 RRAM devices. PHASTRAC will address the most critical issues, namely 1) novel devices for implementing ONN architecture, 2) novel ONN architecture to allow analog sensor data processing, and 3) processing the data efficiently to take appropriate action. This “sensing-to-action” computing approach based on ONN technology will allow energy efficiency improvements.

TU Eindhoven, Netherlands, leads the consortium of four partners. The project partners are Pazmany Peter Catholic University, Hungary, IBM Research Zurich, and Bayerische Motoren Werke AG (BMW), Germany.

With all those exotic materials, it sounds pretty ickky to manufacture. Does this need a new machine?

The ickkyness goes on ... capacitors and varistors ...

IBM has a patent application for ONN:


US2022004876A1 TRAINING OF OSCILLATORY NEURAL NETWORKS 20200702

[0002] Oscillatory neural networks (ONNs) are artificial neural networks which employ a network of coupled oscillators. The oscillators correspond to neurons of a neural network, and the strength of the coupling between pairs of oscillators emulates the network (synaptic) weights. Such networks can be trained to perform correlation tasks, such as image classification and speech recognition, by processing training samples in the network and adapting the matrix of network weights so that the network “learns”, or memorizes, the particular correlations to be detected. Few hardware implementations have been proposed for ONNs, and these networks typically rely on mathematical pre-training of the weights matrix, e.g. via a Hebbian learning algorithm.

The network comprises at least one network layer in which a plurality of electronic oscillators, interconnected via programmable coupling elements storing respective network weights, generate oscillatory signals at time delays dependent on the input signal to propagate the input signal from an input to an output of that layer. The network is adapted to provide a network output signal dependent substantially linearly on phase of oscillatory signals in the last layer of the network. The method includes calculating a network error dependent on the output signal and a desired output for the training sample, and calculating updates for respective network weights by backpropagation of the error such that weight-updates for a network layer are dependent on a vector of time delays at the input to that layer and the calculated error at the output of that layer.

Note the IBM inventors are based in Zurich.

This sounds like spike rate with the added complexity of phase matching, which, added to the inconsistency of capacitors and memRistors sounds messy to implement.

 

[Tothemoon24]

Palo Alto Networks' CTO on why machine learning is revolutionizing SOC performance — VentureBeat

VentureBeat recently sat down (virtually) with Nir Zuk, founder and chief technology officer (CTO) of Palo Alto Networks. Zuk is considered one of the leading technology leaders defining the current and future cybersecurity landscape. Our interview focused on how machine learning is helping...

apple.news

 

[JoMo68]

Puh-leeze!
Why did you have to dig out this misleading comparison once again, although I had already fact-checked it months ago? Well, here we go again:

In the “technical comparison” image you re-posted, Akida gets compared to Dynap-SEL, SynSense’s 2018 neuromorphic chip featuring “1k analog low-power spiking neurons and up to 80k configurable synaptic connections, including 8k synapses with integrated spike-based learning rules.”

Low-Power Neuromorphic Chip DYNAP™-SEL | SynSense

Our next-generation neuromorphic chip, Powering Innovation with 1k Analog Neurons and 80k Configurable Synapses, Including Integrated Spike-Based Learning!

www.synsense.ai

You are evidently aware, though, that BMW are experimenting with SynSense’s fully event-driven neuromorphic vision SoC Speck for their smart cockpit occupant monitoring R&D. So why are you not comparing Akida to Speck instead, as you ought to (although I am not sure whether a direct comparison between AKD1000 and a smart vision processing SoC combining a dynamic vision sensor (DVS) and a neuromorphic processor, actually makes sense?)

https://www.synsense.ai/wp-content/uploads/2023/06/Speck-application-brochures-EN.pdf

View attachment 61625

View attachment 61626

And to be fair, you should also factor in Speck’s competitive price tag at < 7 $ (presumably USD) that could easily tip the scales in SynSense’s favour, when potential customers who don’t mind doing business with a de facto-Chinese company consider Speck’s technical specifications “good enough“ for their envisaged use cases, even though AKD1000 boasts more than three times as many neurons as Speck (but not more than 1000x as many, as your comparison seems to suggest). They may not see the point in paying more for a product that could be described as “over-engineered” for their narrow use cases.

IMO, you are doing BrainChip no favour by cherry-picking a competitor’s far less capable neuromorphic mixed-signal chip for your apples and oranges-comparison. By doing so, you are totally exaggerating the parameter divide between Akida and the competition’s more advanced neuromorphic offerings, eg in SynSense’s case, the fully digital neuromorphic processor Dynap-CNN:

DYNAP™-CNN: Neuromorphic Processor With 1M Spiking eurons

DYNAP™-CNN is world's first fully scalable, event-driven neuromorphic processor with 1M configurable spiking neurons and direct interface with external DVS.

www.synsense.ai

View attachment 61627


What’s the point of unfairly disparaging the competition?
While we shareholders may wish for BrainChip to literally make EVERY sensor smart one day, the commercial reality will be that we will never gain 100% of the global market share. Regardless of any technological superiority.

Puh-leeze!
Why did you have to dig out this misleading comparison once again, although I had already fact-checked it months ago? Well, here we go again:

In the “technical comparison” image you re-posted, Akida gets compared to Dynap-SEL, SynSense’s 2018 neuromorphic chip featuring “1k analog low-power spiking neurons and up to 80k configurable synaptic connections, including 8k synapses with integrated spike-based learning rules.”

Low-Power Neuromorphic Chip DYNAP™-SEL | SynSense

Our next-generation neuromorphic chip, Powering Innovation with 1k Analog Neurons and 80k Configurable Synapses, Including Integrated Spike-Based Learning!

www.synsense.ai

You are evidently aware, though, that BMW are experimenting with SynSense’s fully event-driven neuromorphic vision SoC Speck for their smart cockpit occupant monitoring R&D. So why are you not comparing Akida to Speck instead, as you ought to (although I am not sure whether a direct comparison between AKD1000 and a smart vision processing SoC combining a dynamic vision sensor (DVS) and a neuromorphic processor, actually makes sense?)

https://www.synsense.ai/wp-content/uploads/2023/06/Speck-application-brochures-EN.pdf

View attachment 61625

View attachment 61626

And to be fair, you should also factor in Speck’s competitive price tag at < 7 $ (presumably USD) that could easily tip the scales in SynSense’s favour, when potential customers who don’t mind doing business with a de facto-Chinese company consider Speck’s technical specifications “good enough“ for their envisaged use cases, even though AKD1000 boasts more than three times as many neurons as Speck (but not more than 1000x as many, as your comparison seems to suggest). They may not see the point in paying more for a product that could be described as “over-engineered” for their narrow use cases.

IMO, you are doing BrainChip no favour by cherry-picking a competitor’s far less capable neuromorphic mixed-signal chip for your apples and oranges-comparison. By doing so, you are totally exaggerating the parameter divide between Akida and the competition’s more advanced neuromorphic offerings, eg in SynSense’s case, the fully digital neuromorphic processor Dynap-CNN:

DYNAP™-CNN: Neuromorphic Processor With 1M Spiking eurons

DYNAP™-CNN is world's first fully scalable, event-driven neuromorphic processor with 1M configurable spiking neurons and direct interface with external DVS.

www.synsense.ai

View attachment 61627


What’s the point of unfairly disparaging the competition?
While we shareholders may wish for BrainChip to literally make EVERY sensor smart one day, the commercial reality will be that we will never gain 100% of the global market share. Regardless of any technological superiority.

I do hope that your post is meant to be ‘tongue in cheek’…

 

[Iseki]

Him and DH can only appear when the SP is sub 25 cents.
They've left the braincell in good hands for now (Iseki and Dk)

Shame about the trolling. I would rather this forum be silent for a few hours, between interesting titbits, than fill it with pretend "LTH discussion"

Yeah sure, blah blah

The SP has imploded under the current board, from > $1.30 to 30c in the past few years with not an agreement to show for these years of effort by the staff. A new commercialization board will bring relief, rather than disruption. It is how we support our staff and the founders. Faith based investors, where the faith is based on the personalities of the board, are destroying the company's future because they hardly understand our product.

If Stevens hadn't been given the flick our faith based shareholders would be singing his praises, popsting images of him and the lion and tiger.

Just as Stevens was given his marching orders, if there is no announcement or $2Mill revenue, it is now our chance to stand up and do what's right.

We have the science behind us. The science doesn't rely on personalities. Now stand up and be counted.

 

[wilzy123]

Yeah sure, blah blah

The SP has imploded under the current board, from > $1.30 to 30c in the past few years with not an agreement to show for these years of effort by the staff. A new commercialization board will bring relief, rather than disruption. It is how we support our staff and the founders. Faith based investors, where the faith is based on the personalities of the board, are destroying the company's future because they hardly understand our product.

If Stevens hadn't been given the flick our faith based shareholders would be singing his praises, popsting images of him and the lion and tiger.

Just as Stevens was given his marching orders, if there is no announcement or $2Mill revenue, it is now our chance to stand up and do what's right.

We have the science behind us. The science doesn't rely on personalities. Now stand up and be counted.

This is a woefully lazy effort, even for you Gacy.

Here we see Gazy pumping the following:

1. Decline in share price
2. Criticism of the current board
3. Distrust in faith-based investors
4. Advocacy for change following leadership turnover
5. Urgency for financial results

Gacy doesn't say "why", coz that would be too much effort for lazy Gacy and not suit their lazy agenda.

Keep Praying Buddy

 

[DingoBorat]

"It's like owning a goldmine, where people pay you to mine there and then pay you royalties, with zero capital expenditure, or mining risks."

Well said DB, point well made. The analogy of the goldmine is worth exploring.

What actually happens is that the gold is pre-sold through an off-take arrangement, then that money pays for the mining.

So with us it might be: license the IP, with money down, and we'll guarantee delivery of chips in 6 months. This is much more than we are offering.

The simple fact is that if no-one is signing IP license agreements, then let's invent a few more options for them. It's great that we say we're being flexible, but are we?

That's the whole point of being flexible, Iseki.

An IP licence, isn't the "only" way.

The development and marketing of actual products, with both VVDN and Unigen, are clear evidence of this.

Also, it's easy to say "no-one is signing" but we have no idea at what stage, negotiations with other Companies are.

That's easy to say too, but the confidence exuding from the Company, suggests to me, that there is a lot happening, behind the scenes.

I keep in mind "all" the interactions and partnerships, with other Companies, which are all ongoing and I know, it's just a matter of Time.

I Trust, that the Company know what they are doing, yes they have made strategic mistakes in the past, as all companies do (even behemoths) but we have a very strong team, that are changing tack, when necessary.

When AKIDA was originally "ready to go" it was still ahead of its Time, too much.

The Time is now ripe and yet Time and patience, is still needed, for investors.

 

[Diogenese]

Yeah sure, blah blah

The SP has imploded under the current board, from > $1.30 to 30c in the past few years with not an agreement to show for these years of effort by the staff. A new commercialization board will bring relief, rather than disruption. It is how we support our staff and the founders. Faith based investors, where the faith is based on the personalities of the board, are destroying the company's future because they hardly understand our product.

If Stevens hadn't been given the flick our faith based shareholders would be singing his praises, popsting images of him and the lion and tiger.

Just as Stevens was given his marching orders, if there is no announcement or $2Mill revenue, it is now our chance to stand up and do what's right.

We have the science behind us. The science doesn't rely on personalities. Now stand up and be counted.

Whom do you have in mind for the new commercialization board?

 

[Adam]

P

Him and DH can only appear when the SP is sub 25 cents.
They've left the braincell in good hands for now (Iseki and Dk)

Shame about the trolling. I would rather this forum be silent for a few hours, between interesting titbits, than fill it with pretend "LTH discussion"

Hey Damo..2022,
Respect you opinion and views, as a holder. Many lthers here, are heavily invested and have been long..term..holders and supporters of the product, and still..are. Yeah, well..we are oldies and trust PVDM and the product, but merely, wish, to be told what's happening by the board, to shareholders, and not being told 'its all sunshine and lollipops'. For better ..or for worse, the board needs transparency, to all shareholders. Here endeth, the sermon.(Exciting podcast, man, watch the financials, on potential sales Q2 202X), linked in Developer company XYZ..Appletatateslasamsungxbox.com...

 

[Pmel]

https://www.linkedin.com/feed/updat...date:(V2,urn:li:activity:7189785466844450816)

 

[TheFunkMachine]

There weren't many to start with. They were just for demos. So doubt they will bring much money for revenue lol

If you say that with such confidence then you should know exactly how many they had in stock, otherwise your just guessing at best.

My point is that many business have bought 5-10k in test systems from Brainchip.

Only serious businesses would buy this for demo purposes IMO, and hopefully they like what they see and leads to sales down the line.

Or maybe Intel baught them all to stop othere from buying any

 

[Makeme 2020]

This is a woefully lazy effort, even for you Gacy.

Here we see Gazy pumping the following:

1. Decline in share price
2. Criticism of the current board
3. Distrust in faith-based investors
4. Advocacy for change following leadership turnover
5. Urgency for financial results

Gacy doesn't say "why", coz that would be too much effort for lazy Gacy and not suit their lazy agenda.

Keep Praying Buddy

View attachment 61669

Nice selfie Willy 123.
Love your work DH

 

[supersonic001]

LinkedIn Login, Sign in | LinkedIn

Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.

www.linkedin.com

View attachment 61631

Another really nice and memorable image, but what has the Akida Edge AI Box got to do with “leading engineering innovation in the automotive sector”?

Cars arranged in BRN logo format, chapeau lads ….. on brand every time, consistently!

 

[Berlinforever]

Data centres and Akida? Here a comprehensive review.

https://arxiv.org/pdf/2402.02521

 

[Berlinforever]

This BrainChip presentation got picked up by others too?

TENN: A highly efficient transformer replacement for edge and event processing
M Anthony Lewis, Yan Ru Pei and Olivier Coenen
April 25, 2024

https://flagship.kip.uni-heidelberg.de/jss/HBPm?m=displayPresentation&mI=257&mEID=9439

 

[Diogenese]

This BrainChip presentation got picked up by others too?

TENN: A highly efficient transformer replacement for edge and event processing
M Anthony Lewis, Yan Ru Pei and Olivier Coenen
April 25, 2024

https://flagship.kip.uni-heidelberg.de/jss/HBPm?m=displayPresentation&mI=257&mEID=9439

Good lord!

Learnt about Chebyshev polynomials 50 years ago - didn't think you could actually use them!

 

[Fullmoonfever]

If not posted....McKinsey Eelectronics liking what we're doing highlighting the recent Microchip boards. Nice that one of their preferred partners is Valeo.

MANUFACTURERS | Mckinsey Electronics

A wide range of active and passive components. From MOSFETs, resistors, and sensors to FPGAs and microprocessors, we cater to your circuit design needs with our genuine electronic components and modules.

www.mckinsey-electronics.com

McKinsey Electronics on LinkedIn: #brainchip #akida #mckinseyelectronics

BrainChip recently unveiled its Akida neuromorphic processor on Microchip's embedded platform, using SAMv71 Ultra and SAMA7G54-EK boards. These neuromorphic…

www.linkedin.com

McKinsey Electronics
1,555 followers
1w

BrainChip recently unveiled its Akida neuromorphic processor on Microchip's embedded platform, using SAMv71 Ultra and SAMA7G54-EK boards. These neuromorphic computing systems aim to perform parallel and distributed processing, imitating the structure and functionality of the human brain. It functions on an event-based principle, staying inactive until triggered, leading to decreased power consumption. This board highlights Akida's efficiency with a 32-bit microprocessor unit and a TSMC 28nm technology, focusing on always-on machine learning tasks like keyword spotting and visual wake words. This SoC implements 80 NPUs, offering on-chip learning and a complete machine-learning framework. The second-generation Akida platform supports the intelligence chip market, integrating AI accelerators for IoT and high-performance RF applications. It boasts event-based computing, scalability, and customizable AI neural processing capabilities. BrainChip's software development ecosystem supports seamless neural network creation, training and testing. #BrainChip #Akida #McKinseyElectronics

•