BRN Discussion Ongoing

[Boab]

The very favourable Forbes article is now on the BrainChip website under PR
Enjoy if you haven't already had a chance to see it.

Real-Time AI At The Edge May Require A New Network Solution

Real-Time AI At The Edge May Require A New Network Solution AI has the power to change every electronic platform, but what works in the data center may not work in an industrial edge platform like a security camera, robotic arm, or even a vehicle. There is no one-size-fits-all for edge AI...

brainchip.com

Gotta love this bit.
"Additionally, there are no other off-the-shelf solutions that can provide comparable low-power, high-accuracy, and real-time AI processing".

 

[mcm]

I’m

Had a sneaky look back on HC. still a cesspit!... But, regardless, it is more so challenging everyday to defend the position here that we are not at any risk of being overtaken. Currently, the situation seems as likely as having a contract or actual bloody revenue announce, where is the management? IMO

Pretty much the only posts on the Crapper worth reading are those by Fact Finder and rayz. I much appreciate their efforts.

 

[Tothemoon24]

Northrop Grumman Australia and Spiral Blue team up for missile defence​

6 August 2024



A new iLAuNCH Trailblazer project aims to strengthen defence research in Australia by maturing a full stack missile detection capability.

Northrop Grumman Australia and Spiral Blue are working together on the project.




"We are bringing together Australian knowhow to build an advanced space surveillance system for missile defence tailored for deployment on small satellites,” said iLAuNCH Trailblazer Executive Director, Darin Lovett.

This mission aspires to cultivate a new sovereign capability for hypersonic missile detection through the fusion of cutting-edge infrared detection technology and industry-leading space-based artificial intelligence (AI) and machine learning (ML) data processing techniques.

“Through mission engineering, we plan to rapidly progress promising future operational concepts up the TRL ladder. We will catalyse the expertise in hypersonic vehicle signatures, space-based detection technologies, and the use of AI/ML techniques for threat identification from our different project partners to inform the development of a discriminating capability to the end user,” said Northrop Grumman Australia’s Technology Outreach Manager, Dr Dushy Tissa.

Anticipated outcomes encompass hypersonic vehicle radiance and trajectory modelling, comprehensive design, and rigorous analysis of an infrared electro-optical system, by leveraging state-of-the-art infrared detectors and space edge AI/ML hardware and algorithms for on-board event detection.

 

[Getupthere]

MSI MS-C906 ultra-low-power fanless mini PC Edge AI — Geeky Gadgets

The MSI MS-C906 is a groundbreaking ultra-low-power fanless box IPC designed to meet the demanding requirements of edge AI applications.

apple.news

 

[7für7]

What do you think about this Sony/honda car? I know they collaborate with Microsoft azure.. but you know? Since it’s just a concept car and they plan to bring the car by 2026… there is still time fore some changes right?

Sony? Do you read this? Smart changes… cost effective changes.. both cloudfree and cloudbased available changes available and so on…

AFEELA - EV | ソニー・ホンダモビリティ

AFEELA - 知性を持った存在として人がモビリティを感じること。新しいテクノロジーでモビリティが人や社会を感じること。そんな感じあえる関係性を新たに創り、育てていく。AFEELA。進化はもう始まっている。

www.shm-afeela.com

 

[Fullmoonfever]

Nice find Tech,

I like this bit ...

[0109] The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array), an ASIC (application specific integrated circuit) a neuromorphic research chip, such as Intel's Loihi chip, or a neural network processor, such as BrainChip's Akida™ chip.

Snap.

Accenture just had another patent published in May too for gesture via neuromorphic although doesn't have Gallo on this one.

https://globaldossier.uspto.gov/result/application/US/17992205/101403

 

[IloveLamp]

Haven't listened yet.....

https://www.alcorn.law/podcast/sap187/

 

[Pom down under]

375 Principal Project Engineer jobs in United States

Today's top 375 Principal Project Engineer jobs in United States. Leverage your professional network, and get hired. New Principal Project Engineer jobs added daily.

www.linkedin.com

Job description​

Who We Are:

Join Our Team! The Tactical Aerospace team is a premier supplier for avionics and aerospace technology for new and legacy DoD systems. If you like avionics, radar systems, or even supporting EW & SIGINT systems, Tactical Aerospace is the place to be. Come Join Us! This position has the option of being a fully remote in the following states: AL, AZ, AR, Dis. of Columbia, FL, GA, MA, MD, MI, MN, NH, NJ, NC, OH, OK, OR, TX, UT, VA, WV only. Alternate work locations are San Antonio, Tx.

Objectives of this Role:

• This role is intended to be a lead over Neuromorphic/Cognitive AI research and development team and will be driving strategies and implementations of our AI solutions to meet our customers’ expectations.
• Lead the development, machine learning (ML), and test of AI as applied to Systems, UAS, Avionics, EW, and/or aerospace subsystems.
• Lead the AI team to create, and implement AI technologies/functionality and deployment strategies
• Direct staff in the performance of Literature reviews, interface with academic institutions, lead the development of proposals and lead the implementation and deployment of those systems.
• Lead a development team in code development (python, C), provide AI training to the internal SwRI Staff, lead the AI test group, implement algorithms, and perform various analysis.
• Lead the AI team in the implementation of Spiking Neural Network (SNN) techniques as well as Generation 2 AI.
  

Daily and Monthly Responsibilities:

• Develop Solutions for AI systems and embedded aerospace/avionics systems and subsystems.
• Will work on 2nd and 3rd Gen AI systems (Cognitive & Neuromorphic AI).
• Develop Solutions for neuromorphic systems, EW, SigInt, Situational Awareness, Drones (UAS/UAV), Avionics, AI/ML sensor correlation/fusion, etc.
• Perform Data Science, Data Flow/Analysis duties, provide simulations, and integrate onto hardware.
• Support business development activities.
• Will also support non-AI programs.

SwRI awarded $35.7 million to support cryptologic systems for U.S. Navy

Electronic warfare technology intercepts signals of interest during naval missions

www.swri.org

 

[davidfitz]

Interesting news lately about Brainchip

 

[7für7]

Interesting news lately about Brainchip

View attachment 67652

It’s not his fault actually… the media is just too incompetent to name it as it should be called! NEURALINK!! Elon, neuralink and brainchip…. What is that? this is not even helpful for brainchip to be honest.

 

[Gazzafish]

Top Neuromorphic Computing Stocks for 2024: Ranked by Pure-Play Focus

The human brain consumes about 20 watts of power, which is roughly equivalent to a dim light bulb. Despite this low power consumption, it can perform an estimated 10^16 operations per second. In contrast, a high-end Nvidia GPU consumes about 400 watts of power to perform a “measly” 10^14 to...

exoswan.com

Extract:-
“

Tier 1: Pure-Play Neuromorphic Computing Stocks​

The pure-play neuromorphic computing stocks represent the cutting edge of this nascent industry. These companies stake their entire business model on the potential of these brain-inspired chips. While this focus creates higher risk, it also offers the most direct exposure for investors bullish on neuromorphic computing. With only one public company currently in this tier, it underscores just how early we are in the neuromorphic computing market.

BrainChip Holdings (ASX: BRN)​

BrainChip Holdings (ASX: BRN) is a first-mover in commercial neuromorphic computing, with a focus on energy-efficient edge AI.

Australia-based BrainChip is a pioneer in commercializing neuromorphic computing, focusing on edge AI solutions. The company has developed an Edge AI platform that combines innovative silicon IP, software, and machine learning. This platform includes the Akida neuromorphic processor. Akida is designed to process information in a way that mimics the human brain from a fundamental hardware level. This “imitation” goes beyond the deep neural networks used in today’s AI models.

Brainship enjoys first-mover advantage in commercial neuromorphic computing. The company’s technology has several unique features, including microwatt power consumption and on-chip learning, while being able to support standard machine learning workflows. In fact, it offers a claimed 5-10x improvement in performance-per-watt over traditional AI accelerators. This would make the Akida chip ideal for battery-powered devices, edge computing, and in-sensor intelligence.

The company is pursuing a flexible business model centered on high-margin IP licensing. This strategy involves upfront license fees and ongoing royalties, which could provide steady revenue as adoption grows. BrainChip’s intellectual property portfolio includes 17 granted patents and 30 pending patents. The company’s team consists of 80% engineers, with 15% holding PhDs from leading AI research programs. BrainChip is also building partnerships with system integrators, including MegaChips, Prophesee, and SiFive.”

 

[Evermont]

Intel Foundry Achieves Major Milestones​

Intel 18A powered on and healthy, on track for next-gen client and server chip production next year.

BrainChip is an IP Partner of IFS. Worth reading the second link as well.

Intel Foundry Achieves Major Milestones

Intel 18A powered on and healthy, on track for next-gen client and server chip production next year.

www.intel.com

Kevin O’Buckley Talks Progress on Intel 18A

New Foundry Services leader reveals why he came to Intel, excitement for the foundry business and why Intel 18A matters to the future of computing.

www.intel.com

 

[Smoothsailing]

Regarding this Intel 18A platform if I am reading this correctly it will be the customer whom chooses their particular design and if it is to include BRN Akida , is that correct?.

 

[Diogenese]

Top Neuromorphic Computing Stocks for 2024: Ranked by Pure-Play Focus

The human brain consumes about 20 watts of power, which is roughly equivalent to a dim light bulb. Despite this low power consumption, it can perform an estimated 10^16 operations per second. In contrast, a high-end Nvidia GPU consumes about 400 watts of power to perform a “measly” 10^14 to...

exoswan.com

Extract:-
“

Tier 1: Pure-Play Neuromorphic Computing Stocks​

The pure-play neuromorphic computing stocks represent the cutting edge of this nascent industry. These companies stake their entire business model on the potential of these brain-inspired chips. While this focus creates higher risk, it also offers the most direct exposure for investors bullish on neuromorphic computing. With only one public company currently in this tier, it underscores just how early we are in the neuromorphic computing market.

BrainChip Holdings (ASX: BRN)​

BrainChip Holdings (ASX: BRN) is a first-mover in commercial neuromorphic computing, with a focus on energy-efficient edge AI.

Australia-based BrainChip is a pioneer in commercializing neuromorphic computing, focusing on edge AI solutions. The company has developed an Edge AI platform that combines innovative silicon IP, software, and machine learning. This platform includes the Akida neuromorphic processor. Akida is designed to process information in a way that mimics the human brain from a fundamental hardware level. This “imitation” goes beyond the deep neural networks used in today’s AI models.

Brainship enjoys first-mover advantage in commercial neuromorphic computing. The company’s technology has several unique features, including microwatt power consumption and on-chip learning, while being able to support standard machine learning workflows. In fact, it offers a claimed 5-10x improvement in performance-per-watt over traditional AI accelerators. This would make the Akida chip ideal for battery-powered devices, edge computing, and in-sensor intelligence.

The company is pursuing a flexible business model centered on high-margin IP licensing. This strategy involves upfront license fees and ongoing royalties, which could provide steady revenue as adoption grows. BrainChip’s intellectual property portfolio includes 17 granted patents and 30 pending patents. The company’s team consists of 80% engineers, with 15% holding PhDs from leading AI research programs. BrainChip is also building partnerships with system integrators, including MegaChips, Prophesee, and SiFive.”

Hi Gazza,

Thanks for this. We've seen a couple of articles along this line, which allow me to maintain my illusion about software as a product.

As you may or may not know, I've been posting about the possibility of our EAPs using Akida simulation software, particularly after the emergence of Akida 2 + TeNNs more than 30 months ago, with Valeo and MB using software for signal processing. This sentence from the article again adds more grist to that rumor mill:

"The company has developed an Edge AI platform that combines innovative silicon IP, software, and machine learning."

To repeat myself, no potential user would commit to Akida 2 in silicon while the tech was in a state of flux. The use of software AI is not so problematic in ICEs as it is in EVs, but from what we've heard about TeNNs, the power and latency could be tolerated in EVs using TeNNs in software. Of course, the software would sensibly include the full Akida 2 simulation including TeNNs, or TeNNs could be used on its own. Software can be readily updated as new developments are implemented, whereas silicon is set in stone.

It's been several months since Anil announced the proposed tapeout of Akida 2, which suggests that the development had reached a satisfactory plateau of stability sufficient for the company to commit to silicon. The tapeout was subsequently "delegated" to a mysterious "other" - the rest is silence.

The SPP talks about developing a cloud-based FPGA demonstration setup, again, not a tapeout. This would be a cheaper way to obtain customer feedback than taping out and making a batch of "engineering sample" chips.

Presumably the mysterious "other" would want to be in on the results of the cloud feedback before going to silicon.

Does this mean that we need to wait for the cloud FPGA venture to provide meaningful results before the tapeout can be implemented? - sigh!

Does it follow that BRN will become a software provider, at least in the short term?

 

[MadMayHam | 合氣道]

Hi Gazza,

Thanks for this. We've seen a couple of articles along this line, which allow me to maintain my illusion about software as a product.

As you may or may not know, I've been posting about the possibility of our EAPs using Akida simulation software, particularly after the emergence of Akida 2 + TeNNs more than 30 months ago, with Valeo and MB using software for signal processing. This sentence from the article again adds more grist to that rumor mill:

"The company has developed an Edge AI platform that combines innovative silicon IP, software, and machine learning."

To repeat myself, no potential user would commit to Akida 2 in silicon while the tech was in a state of flux. The use of software AI is not so problematic in ICEs as it is in EVs, but from what we've heard about TeNNs, the power and latency could be tolerated in EVs using TeNNs in software. Of course, the software would sensibly include the full Akida 2 simulation including TeNNs, or TeNNs could be used on its own. Software can be readily updated as new developments are implemented, whereas silicon is set in stone.

It's been several months since Anil announced the proposed tapeout of Akida 2, which suggests that the development had reached a satisfactory plateau of stability sufficient for the company to commit to silicon. The tapeout was subsequently "delegated" to a mysterious "other" - the rest is silence.

The SPP talks about developing a cloud-based FPGA demonstration setup, again, not a tapeout. This would be a cheaper way to obtain customer feedback than taping out and making a batch of "engineering sample" chips.

Presumably the mysterious "other" would want to be in on the results of the cloud feedback before going to silicon.

Does this mean that we need to wait for the cloud FPGA venture to provide meaningful results before the tapeout can be implemented? - sigh!

Does it follow that BRN will become a software provider, at least in the short term?

@Diogenese what do you mean by "The tapeout was subsequently "delegated" to a mysterious "other" - " ?

 

[Frangipani]

Researchers at UC Irvine’s Cognitive Anteater Robotics Laboratory (CARL), led by Jeffrey Krichmar, have been experimenting with AKD1000:

CARL: Home Page

CARLsim: GPU-accelerated Spiking Neural Network (SNN) simulator

GitHub - UCI-CARL/CARLsimPP

Contribute to UCI-CARL/CARLsimPP development by creating an account on GitHub.

github.com

https://arxiv.org/pdf/2404.08726

 

[FJ-215]

Hi Gazza,

Thanks for this. We've seen a couple of articles along this line, which allow me to maintain my illusion about software as a product.

As you may or may not know, I've been posting about the possibility of our EAPs using Akida simulation software, particularly after the emergence of Akida 2 + TeNNs more than 30 months ago, with Valeo and MB using software for signal processing. This sentence from the article again adds more grist to that rumor mill:

"The company has developed an Edge AI platform that combines innovative silicon IP, software, and machine learning."

To repeat myself, no potential user would commit to Akida 2 in silicon while the tech was in a state of flux. The use of software AI is not so problematic in ICEs as it is in EVs, but from what we've heard about TeNNs, the power and latency could be tolerated in EVs using TeNNs in software. Of course, the software would sensibly include the full Akida 2 simulation including TeNNs, or TeNNs could be used on its own. Software can be readily updated as new developments are implemented, whereas silicon is set in stone.

It's been several months since Anil announced the proposed tapeout of Akida 2, which suggests that the development had reached a satisfactory plateau of stability sufficient for the company to commit to silicon. The tapeout was subsequently "delegated" to a mysterious "other" - the rest is silence.

The SPP talks about developing a cloud-based FPGA demonstration setup, again, not a tapeout. This would be a cheaper way to obtain customer feedback than taping out and making a batch of "engineering sample" chips.

Presumably the mysterious "other" would want to be in on the results of the cloud feedback before going to silicon.

Does this mean that we need to wait for the cloud FPGA venture to provide meaningful results before the tapeout can be implemented? - sigh!

Does it follow that BRN will become a software provider, at least in the short term?

Hi @Diogenese,

One question I would have for Sean, is, how do we provide yield numbers for Akida 2 without proving it out in silicon? Yes, we have AKD 1000/1500 from two different foundries but no commercial runs to speak of.

If I remember the ARM history correctly, they were on their 4th or 5th run of commercial chips before going down the IP route that BRN is trying to copy.

 

[FiveBucks]

Hi @Diogenese,

One question I would have for Sean, is, how do we provide yield numbers for Akida 2 without proving it out in silicon? Yes, we have AKD 1000/1500 from two different foundries but no commercial runs to speak of.

If I remember the ARM history correctly, they were on their 4th or 5th run of commercial chips before going down the IP route that BRN is trying to copy.

Did we jump the gun by going the IP route?

Or are we ahead of the game?

 

[FJ-215]

Did we jump the gun by going the IP route?

Or are we ahead of the game?

Who knows, maybe a mix of the two.

 

[Frangipani]

Researchers at UC Irvine’s Cognitive Anteater Robotics Laboratory (CARL), led by Jeffrey Krichmar, have been experimenting with AKD1000:








View attachment 67694

CARL: Home Page

View attachment 67690


View attachment 67691

View attachment 67692

CARLsim: GPU-accelerated Spiking Neural Network (SNN) simulator

View attachment 67693

GitHub - UCI-CARL/CARLsimPP

Contribute to UCI-CARL/CARLsimPP development by creating an account on GitHub.

github.com

View attachment 67695

https://arxiv.org/pdf/2404.08726

https://arxiv.org/pdf/2404.08726

This is the paper I linked in my previous post, co-authored by Lars Niedermeier, a Zurich-based IT consultant, and the above-mentioned Jeff Krichmar from UC Irvine.

The two of them co-authored three papers in recent years, including one in 2022 with another UC Irvine professor and member of the CARL team, Nikil Dutt (https://ics.uci.edu/~dutt/) as well as Anup Das from Drexel University, whose endorsement of Akida is quoted on the BrainChip website:

Lars Niedermeier’s and Jeff Krichmar’s April 2024 publication on CARLsim++ (which does not mention Akida) ends with the following conclusion and the acknowledgement that their work was supported by the Air Force Office of Scientific Research - the funding has been going on at least since 2022 -

UCI-led team uses robots to study brain’s ability to switch between first-person experience and global map

www.socsci.uci.edu

and a UCI Beall Applied Innovation Proof of Product Award (https://innovation.uci.edu/pop/)

and they also thank the regional NSF I-Corps (= Innovation Corps) for valuable insights.

Their use of an E-Puck robot (https://en.m.wikipedia.org/wiki/E-puck_mobile_robot) for their work reminded me of our CTO’s address at the AGM in May, during which he envisioned the following object (from 22:44 min):

“Imagine a compact device similar in size to a hockey puck that combines speech recognition, LLMs and an intelligent agent capable of controlling your home’s lighting, assisting with home repairs and much more. All without needing constant connectivity or having to worry about privacy and security concerns, a major barrier to adaptation, particularly in industrial settings.”

Possibly something in the works here?

The version the two authors were envisioning in their April 2024 paper is, however, conceptualised as being available as a cloud service:

“We plan a hybrid approach to large language models available as cloud service for processing of voice and text to speech.”


The authors gave a tutorial on CARLsim++ at NICE 2024, where our CTO Tony Lewis was also presenting. Maybe they had a fruitful discussion at that conference in La Jolla, which resulted in UC Irvine’s Cognitive Anteater Robotics Laboratory (CARL) team experimenting with AKD1000, as evidenced in the video uploaded a couple of hours ago that I shared in my previous post?

NICE 2024 - Agenda

GCtronic

www.gctronic.com