BRN Discussion Ongoing

[buena suerte :-)]

80's teen too! I cringe when I think of my favourite songs. My husband LOVES to embarass me - he had way cooler taste in music. I seemed to have a penchant for the Stock Aitken Waterman production line And what about the shoulder pads!!

Love it good old S..A..and Waterman

 

[7für7]

Some people need obviously money for the weekend party ‍

 

[Diogenese]

So now it's got a fan?

The NXP i.MX 8M Plus Quad SOC must have been running hot.
https://shop.brainchipinc.com/products/akida™-edge-ai-box

 

[Fullmoonfever]

Thought I’d remembered someone asking about this but in a previous presentation the information wasn’t blank. I wonder why they left it blank in the last presentation?




View attachment 59656 View attachment 59657

Appears to be a diff basket if you look at the general heading of the tables imo.

The first one more of a general competitor analysis of neuromorphic or aligned processors eg Loihi (research), True North (not edge) etc.

The second presso table is "edge" competitor aligned by the looks.

Regardless of how we want Akida to be the front runner, there are diff processors appearing at the edge or fog or hybrid edge/cloud and they will be competitors whether we like it or not.

Part of why I wanted to know who they were so we could individually assess the competitors and do our own DD / research.

 

[Fullmoonfever]

So now it's got a fan?

View attachment 59682


The NXP i.MX 8M Plus Quad SOC must have been running hot.
https://shop.brainchipinc.com/products/akida™-edge-ai-box

Unless it's a generic casing / form factor and fan opening is there for other board set ups internally that do run hot?

Didn't they say there are other players who can be used within the box depending on what client wanted?

 

[Diogenese]

www.gizchina.com/2024/03/21/samsung-mach-1-ai-chip/amp/


"Samsung’s entry into the AI chip market with the Mach-1 does not directly aim at competing with ultra-high-end solutions like Nvidia’s H100 or B100 series. Instead, Samsung is focusing on developing an AI inference accelerator based on an application-specific integrated circuit (ASIC) that excels in edge computing applications. The company’s emphasis on low power consumption, minimal dimensions, and cost-effectiveness aligns with the growing demand for efficient AI solutions across various industries."

Happy as Larry

Hi TD,

"Samsung’s innovative approach with the Mach-1 involves utilizing LPDDR memory, a unique feature that significantly reduces memory bandwidth requirements for inference to about 0.125x when compared to existing designs. This strategic use of LPDDR memory positions the Mach-1 as a lightweight AI chip that does not rely on high bandwidth memory (HBM), which is currently scarce and expensive in the market."

This is not one of our chickens:

US2024086153A1 MULTI-BIT ACCUMULATOR AND IN-MEMORY COMPUTING PROCESSOR WITH SAME 20220914

A multi-bit accumulator includes 1-bit Wallace trees each configured to perform an add operation on single-bit input data, tristate logic circuits each configured to output a result of the add operation of the 1-bit Wallace trees according to an enable signal provided to the tristate logic circuits, and a shift-adder configured to perform an accumulation operation on the result of the add operation of the 1-bit Wallace trees by a shift operation based on a clock signal.
[0171] IMC [in-memory compute] may accelerate a matrix operation and/or an MAC operation that performs addition of multiple multiplications at a time, which may be general for artificial intelligence (AI) learning and inference. The MAC operation for the learning and inference of the neural network 1210 may be performed through a memory array, and the memory array may include bit cells of a memory device in the IMC macro.

 

[AARONASX]

So now it's got a fan?

View attachment 59682


The NXP i.MX 8M Plus Quad SOC must have been running hot.
https://shop.brainchipinc.com/products/akida™-edge-ai-box

Might not be there for the chip itself, it could be box fan to assist in ventilation for other components that could create heat. eg transformers, Resistors. Capacitors. Inductors or. power supply etc

 

[Fullmoonfever]

Know we know we are attending but got our booth and outine of what we doing there.

2024 Exhibitor Directory

We’re Proud of Our 2024 Sponsors, Exhibitors and Partners—with more to come! We’re proud of the first-rate exhibitors at the Embedded Vision Summit. They’re focused on practical, deployable computer vision and edge AI. Our exhibitors bring to the conference some […]

embeddedvisionsummit.com

We're Proud of Our 2024 Sponsors, Exhibitors and Partners—with more to come!​

We’re proud of the first-rate exhibitors at the Embedded Vision Summit. They’re focused on practical, deployable computer vision and edge AI. Our exhibitors bring to the conference some quite remarkable technologies and solutions and provide fascinating demos. Connect with their technical experts and check out their newest technologies in their booths!​

BrainChipGoldExhibitorsEnabling Technologies Track: Product Presentations​

BrainChip will demonstrate a range of use cases based on Akida, an ultra-low-power, fully digital, event-based neuromorphic processor that targets advanced security, industrial, automotive and augmented reality applications. We will spotlight the Akida Edge AI Box, a compact box designed for video analytics, facial recognition and object detection, and can extend intelligent processing capabilities that integrate inputs from various other sensors.

Booth: #618

 

[Bravo]

Hey Brain Fam,

Re: Homomorphic Encryption

I think it might be a wee bit interesting to look further into the connection between Intel, Microsoft and DARPA (and potentially us).

WARNING: Please avert you eyes if you don't appreciate the ancient art of dot-joining!

https://www.eetimes.com/darpa-funds-asic-accelerators-for-fully-homomorphic-encryption/

1. Here's an extract from an Intel blog dated 8 March 2021 describing INTEL collaborating with MICROSOFT on the the DARPA DPRIVE program (more info in the above link) with plans to design an application specific ASIC accelerator for Homomorphic Encryption. The DPRIVE program is set to wrap up around September 2024.

Intel to Collaborate with Microsoft on DARPA Program

Intel announced that it has signed an agreement with Defense Advanced Research Projects Agency to perform in its Data Protection in Virtual Environments program.

www.intel.com

2. Secondly, we know that we joined the Intel Foundry Services IP Alliance in December 2022.

https://brainchip.com/brainchip-joins-intel-foundry-services-to-advance-neuromorphic-ai-at-the-edge/



3. Thirdly, I stumbled over this research paper

A Homomorphic Encryption Framework for Privacy-Preserving Spiking Neural Networks​

published October 2023 in which also happens to references the DARPA DPRIVE Program. It concludes by describing SNN's as being a crucial factor in the development of future private and secure networks and also states that this work demonstrates the applicability of Homomorphic Encryption schemes with SNN's.



a) Extract

b) Extract

A Homomorphic Encryption Framework for Privacy-Preserving Spiking Neural Networks

Machine learning (ML) is widely used today, especially through deep neural networks (DNNs); however, increasing computational load and resource requirements have led to cloud-based solutions. To address this problem, a new generation of networks has emerged called spiking neural networks (SNNs)...

www.mdpi.com

3. Fourthly, here is a recent announcement whereby Microsoft has chosen Intel as a chipmaking partner.

Intel and Microsoft just announced a huge collaboration | Digital Trends

Microsoft will be leveraging Intel's 18A technology for the production of an undisclosed chip.

www.digitaltrends.com

As a side note, BrainChips involvement in the IP Alliance would presumably give Microsoft the ability to access our IP.

Make of this what you will.

 

[TheDrooben]

Hi TD,

"Samsung’s innovative approach with the Mach-1 involves utilizing LPDDR memory, a unique feature that significantly reduces memory bandwidth requirements for inference to about 0.125x when compared to existing designs. This strategic use of LPDDR memory positions the Mach-1 as a lightweight AI chip that does not rely on high bandwidth memory (HBM), which is currently scarce and expensive in the market."

This is not one of our chickens:

US2024086153A1 MULTI-BIT ACCUMULATOR AND IN-MEMORY COMPUTING PROCESSOR WITH SAME 20220914

A multi-bit accumulator includes 1-bit Wallace trees each configured to perform an add operation on single-bit input data, tristate logic circuits each configured to output a result of the add operation of the 1-bit Wallace trees according to an enable signal provided to the tristate logic circuits, and a shift-adder configured to perform an accumulation operation on the result of the add operation of the 1-bit Wallace trees by a shift operation based on a clock signal.
[0171] IMC [in-memory compute] may accelerate a matrix operation and/or an MAC operation that performs addition of multiple multiplications at a time, which may be general for artificial intelligence (AI) learning and inference. The MAC operation for the learning and inference of the neural network 1210 may be performed through a memory array, and the memory array may include bit cells of a memory device in the IMC macro.

Thanks Dio.....

 

[davidfitz]

Leaked docs hint Google may use SiFive RISC-V cores in next-gen TPUs​

MSN

This event may shed some more light on what Google has planned?

AI, Android and more, Google I/O is coming May 14th

Google have announced the dates for their developer conference, Google I/O this week, with the three day conference set to kick off on May 14th in the US.

eftm.com

Google have announced the dates for their developer conference, Google I/O this week, with the three day conference set to kick off on May 14th in the US.

The event will be streamed ‘in front of a small live audience’ on May 14 in the US, which will be in the early hours of May 15th here in Australia. The conference will begin with the usual Keynote speech where we’ll get an overview of what to expect from Google over the next 12 months.

The focus of Google I/O will become clearer as developer sessions are announced closer to the event, but we expect to see an intense focus on AI, more Android announcements and of course Search, Maps and more.

While Google has previously announced hardware on-stage at I/O, we don’t expect to see more on the Pixel 8a – but we could be in for a surprise. We may also see more on the slate of upcoming rumoured hardware which includes the Pixel Tablet 2, Pixel Watch 3, and maybe even Pixel Buds Pro gen 2. We also expect Google to tease the Pixel 9 and Pixel 9 Pro which should be unveiled in October.

You can find more about Google I/O on their landing page, but we’ll be bringing you all the latest news as it’s announced on May 15th.

 

[Slade]

Know we know we are attending but got our booth and outine of what we doing there.

2024 Exhibitor Directory

We’re Proud of Our 2024 Sponsors, Exhibitors and Partners—with more to come! We’re proud of the first-rate exhibitors at the Embedded Vision Summit. They’re focused on practical, deployable computer vision and edge AI. Our exhibitors bring to the conference some […]

embeddedvisionsummit.com

We're Proud of Our 2024 Sponsors, Exhibitors and Partners—with more to come!​

We’re proud of the first-rate exhibitors at the Embedded Vision Summit. They’re focused on practical, deployable computer vision and edge AI. Our exhibitors bring to the conference some quite remarkable technologies and solutions and provide fascinating demos. Connect with their technical experts and check out their newest technologies in their booths!​

BrainChipGoldExhibitorsEnabling Technologies Track: Product Presentations​

BrainChip will demonstrate a range of use cases based on Akida, an ultra-low-power, fully digital, event-based neuromorphic processor that targets advanced security, industrial, automotive and augmented reality applications. We will spotlight the Akida Edge AI Box, a compact box designed for video analytics, facial recognition and object detection, and can extend intelligent processing capabilities that integrate inputs from various other sensors.

Booth: #618

Thanks for sharing FMF. A use case for augmented reality applications. Is this new?

 

[Fullmoonfever]

Thanks for sharing FMF. A use case for augmented reality applications. Is this new?

Nice pick up Slade.

I missed those couple of words seeing more re edge box demoing again.

 

[Slade]

Nice pick up Slade.

I missed those couple of words seeing more re edge box demoing again.

It’s certainly timely.

Integrating Augmented Reality with Neuromorphic Computing for Enhanced Cognitive Training​

Aardvark Infinity
·
Follow
3 min read
·
Mar 9, 2024

Listen
Share
Author: Aardvark Infinity | www.aardvarkinfinity.com

Published on Mar 8, 2024 · 4 min read

Photo by Lucrezia Carnelos on Unsplash
In the ever-evolving landscape of technology, the fusion of augmented reality (AR) and neuromorphic computing represents a groundbreaking stride toward the future of cognitive training and rehabilitation. Aardvark Infinity is at the forefront of this innovation with its latest project, “Augmented Cognition: Neuromorphic AR Systems for Cognitive Enhancement (AC-NAReS).” This pioneering initiative aims to leverage the unique strengths of neuromorphic computing — computers inspired by the neural structure of the human brain — and AR’s immersive experiences to create personalized cognitive training programs.

### Bridging the Gap between Brain and Machine

Neuromorphic computing mimics the human brain’s neural architecture, offering unparalleled efficiency in processing complex, sensory-driven tasks. By simulating the way neurons communicate and learn, neuromorphic systems can analyze and interpret vast amounts of data more efficiently than traditional computing architectures. This efficiency makes them ideal for applications requiring real-time processing and adaptive learning, such as cognitive training exercises.

Augmented reality, on the other hand, enriches the real world with digital overlays, providing users with immersive, interactive experiences. When combined with neuromorphic computing, AR can deliver highly personalized, adaptive learning environments that respond to the user’s cognitive state and needs in real-time. This synergy offers a promising avenue for cognitive training and rehabilitation, providing a dynamic platform that adjusts to the user’s progress and challenges them appropriately.

### The AC-NAReS Project: A Leap Forward in Cognitive Training

Aardvark Infinity’s AC-NAReS project aims to develop a platform that integrates neuromorphic computing with AR technology to support cognitive training and rehabilitation. The project focuses on several key areas:

1. **Real-time Cognitive State Monitoring:** Using neuromorphic sensors and algorithms, the system continuously assesses the user’s cognitive state, providing immediate feedback to adjust the difficulty and type of cognitive exercises.

2. **Immersive Learning Environments:** AR creates engaging, interactive scenarios that can be tailored to target specific cognitive skills, such as memory, attention, and problem-solving.

3. **Adaptive Learning Pathways:** Leveraging the learning capabilities of neuromorphic computing, the system adapts in real-time to the user’s performance, personalizing the training regimen to optimize cognitive development and recovery.

4. **Accessibility and Portability:** By utilizing wearable AR devices and neuromorphic computing’s efficient processing, the AC-NAReS platform aims to be accessible anywhere, enabling users to engage in cognitive training in the comfort of their homes or while on the move.

### Implications and Future Directions

The integration of augmented reality with neuromorphic computing in cognitive training opens up new possibilities for personalized education, therapeutic interventions, and cognitive enhancement. The AC-NAReS project not only aims to advance our understanding of how technology can support cognitive development but also strives to make these technologies accessible to a broader audience.

As this project progresses, Aardvark Infinity envisions a future where cognitive training and rehabilitation are not only more effective but also more engaging and accessible to individuals across the globe. By pushing the boundaries of what’s possible with AR and neuromorphic computing, we are stepping into a new era of cognitive enhancement that promises to enrich the human experience in profound ways.

Stay tuned for updates on the AC-NAReS project and explore how Aardvark Infinity continues to lead innovation at the intersection of neuroscience, computing, and augmented reality at

 

[Fullmoonfever]

It’s certainly timely.

Integrating Augmented Reality with Neuromorphic Computing for Enhanced Cognitive Training​

Aardvark Infinity
·
Follow
3 min read
·
Mar 9, 2024

Listen
Share
Author: Aardvark Infinity | www.aardvarkinfinity.com

Published on Mar 8, 2024 · 4 min read

Photo by Lucrezia Carnelos on Unsplash
In the ever-evolving landscape of technology, the fusion of augmented reality (AR) and neuromorphic computing represents a groundbreaking stride toward the future of cognitive training and rehabilitation. Aardvark Infinity is at the forefront of this innovation with its latest project, “Augmented Cognition: Neuromorphic AR Systems for Cognitive Enhancement (AC-NAReS).” This pioneering initiative aims to leverage the unique strengths of neuromorphic computing — computers inspired by the neural structure of the human brain — and AR’s immersive experiences to create personalized cognitive training programs.

### Bridging the Gap between Brain and Machine

Neuromorphic computing mimics the human brain’s neural architecture, offering unparalleled efficiency in processing complex, sensory-driven tasks. By simulating the way neurons communicate and learn, neuromorphic systems can analyze and interpret vast amounts of data more efficiently than traditional computing architectures. This efficiency makes them ideal for applications requiring real-time processing and adaptive learning, such as cognitive training exercises.

Augmented reality, on the other hand, enriches the real world with digital overlays, providing users with immersive, interactive experiences. When combined with neuromorphic computing, AR can deliver highly personalized, adaptive learning environments that respond to the user’s cognitive state and needs in real-time. This synergy offers a promising avenue for cognitive training and rehabilitation, providing a dynamic platform that adjusts to the user’s progress and challenges them appropriately.

### The AC-NAReS Project: A Leap Forward in Cognitive Training

Aardvark Infinity’s AC-NAReS project aims to develop a platform that integrates neuromorphic computing with AR technology to support cognitive training and rehabilitation. The project focuses on several key areas:

1. **Real-time Cognitive State Monitoring:** Using neuromorphic sensors and algorithms, the system continuously assesses the user’s cognitive state, providing immediate feedback to adjust the difficulty and type of cognitive exercises.

2. **Immersive Learning Environments:** AR creates engaging, interactive scenarios that can be tailored to target specific cognitive skills, such as memory, attention, and problem-solving.

3. **Adaptive Learning Pathways:** Leveraging the learning capabilities of neuromorphic computing, the system adapts in real-time to the user’s performance, personalizing the training regimen to optimize cognitive development and recovery.

4. **Accessibility and Portability:** By utilizing wearable AR devices and neuromorphic computing’s efficient processing, the AC-NAReS platform aims to be accessible anywhere, enabling users to engage in cognitive training in the comfort of their homes or while on the move.

### Implications and Future Directions

The integration of augmented reality with neuromorphic computing in cognitive training opens up new possibilities for personalized education, therapeutic interventions, and cognitive enhancement. The AC-NAReS project not only aims to advance our understanding of how technology can support cognitive development but also strives to make these technologies accessible to a broader audience.

As this project progresses, Aardvark Infinity envisions a future where cognitive training and rehabilitation are not only more effective but also more engaging and accessible to individuals across the globe. By pushing the boundaries of what’s possible with AR and neuromorphic computing, we are stepping into a new era of cognitive enhancement that promises to enrich the human experience in profound ways.

Stay tuned for updates on the AC-NAReS project and explore how Aardvark Infinity continues to lead innovation at the intersection of neuroscience, computing, and augmented reality at

Geez...we mustn't like the real world much huh

"Augmented reality, on the other hand, enriches the real world with digital overlays, providing users with immersive, interactive experiences."

 

[Fullmoonfever]

It’s certainly timely.

Integrating Augmented Reality with Neuromorphic Computing for Enhanced Cognitive Training​

Aardvark Infinity
·
Follow
3 min read
·
Mar 9, 2024

Listen
Share
Author: Aardvark Infinity | www.aardvarkinfinity.com

Published on Mar 8, 2024 · 4 min read

Photo by Lucrezia Carnelos on Unsplash
In the ever-evolving landscape of technology, the fusion of augmented reality (AR) and neuromorphic computing represents a groundbreaking stride toward the future of cognitive training and rehabilitation. Aardvark Infinity is at the forefront of this innovation with its latest project, “Augmented Cognition: Neuromorphic AR Systems for Cognitive Enhancement (AC-NAReS).” This pioneering initiative aims to leverage the unique strengths of neuromorphic computing — computers inspired by the neural structure of the human brain — and AR’s immersive experiences to create personalized cognitive training programs.

### Bridging the Gap between Brain and Machine

Neuromorphic computing mimics the human brain’s neural architecture, offering unparalleled efficiency in processing complex, sensory-driven tasks. By simulating the way neurons communicate and learn, neuromorphic systems can analyze and interpret vast amounts of data more efficiently than traditional computing architectures. This efficiency makes them ideal for applications requiring real-time processing and adaptive learning, such as cognitive training exercises.

Augmented reality, on the other hand, enriches the real world with digital overlays, providing users with immersive, interactive experiences. When combined with neuromorphic computing, AR can deliver highly personalized, adaptive learning environments that respond to the user’s cognitive state and needs in real-time. This synergy offers a promising avenue for cognitive training and rehabilitation, providing a dynamic platform that adjusts to the user’s progress and challenges them appropriately.

### The AC-NAReS Project: A Leap Forward in Cognitive Training

Aardvark Infinity’s AC-NAReS project aims to develop a platform that integrates neuromorphic computing with AR technology to support cognitive training and rehabilitation. The project focuses on several key areas:

1. **Real-time Cognitive State Monitoring:** Using neuromorphic sensors and algorithms, the system continuously assesses the user’s cognitive state, providing immediate feedback to adjust the difficulty and type of cognitive exercises.

2. **Immersive Learning Environments:** AR creates engaging, interactive scenarios that can be tailored to target specific cognitive skills, such as memory, attention, and problem-solving.

3. **Adaptive Learning Pathways:** Leveraging the learning capabilities of neuromorphic computing, the system adapts in real-time to the user’s performance, personalizing the training regimen to optimize cognitive development and recovery.

4. **Accessibility and Portability:** By utilizing wearable AR devices and neuromorphic computing’s efficient processing, the AC-NAReS platform aims to be accessible anywhere, enabling users to engage in cognitive training in the comfort of their homes or while on the move.

### Implications and Future Directions

The integration of augmented reality with neuromorphic computing in cognitive training opens up new possibilities for personalized education, therapeutic interventions, and cognitive enhancement. The AC-NAReS project not only aims to advance our understanding of how technology can support cognitive development but also strives to make these technologies accessible to a broader audience.

As this project progresses, Aardvark Infinity envisions a future where cognitive training and rehabilitation are not only more effective but also more engaging and accessible to individuals across the globe. By pushing the boundaries of what’s possible with AR and neuromorphic computing, we are stepping into a new era of cognitive enhancement that promises to enrich the human experience in profound ways.

Stay tuned for updates on the AC-NAReS project and explore how Aardvark Infinity continues to lead innovation at the intersection of neuroscience, computing, and augmented reality at

Reminds me of a movie I saw as a young bloke....is this where we're heading?

Glad our name wasn't Brainstorm haha

 

[Fullmoonfever]

Reminds me of a movie I saw as a young bloke....is this where we're heading?

Glad our name wasn't Brainstorm haha

Or the other one I recall watching.

 

[7für7]

Can we expect a price sensitive announcement after closing?

 

[jtardif999]

EDIT: Right you are - they make the Tegra for the nintendo, a partner with MegaChips, so maybe something cam come of that.

Using licensed Arm IP, NVIDIA has designed Grace as a CPU specifically for giant-scale AI and HPC. Coupled with the GPU and DPU, Grace gives us the third foundational technology for computing, and the ability to re-architect the data center to advance AI. NVIDIA is now a three-chip company.”

 

[Iseki]

My letter to Sean:

Dear Sean,

Looking forward to meeting you at the up-coming AGM. Boy did we have a great time last year!

It's so good that you are excited about the year ahead, and that you have taken some difficult decision this year. I'm so glad that rotter Stevens was given the boot, let me tell you (even though he was a good dancer - what was it? The limbo?)

But this brings me to the reason for my email. We can only okay remuneration based on achieved results and not the effort expended.

These are difficult times, and well, we all need to tighten our belts. I realize that you have a grand plan to sell only IP and not chips, as this would maximize our profit down the tack. But we are not down the track yet, as I'm sure you'll agree.

If only you had decided to manufacture a small batch of our most powerful Akida2 variant, we might have signed off on a new licensing agreement that would have paid for the manufacturing, and a big pile of remuneration for you. But you stuck to your guns and stayed true to your word!

Your firm approach is commendable, except that your remuneration bonus will need to be delayed until your planned super-revenue can be realized.

Sadly we can not postpone your remuneration bonus, unless this is something that you organize. We can only vote on what you put before us.

Thus it is with profoud regret that we must vote down this year's bonus even though it will cause some chaos in the office and on the ASX. But then shares do what shares will do. But just imagine what we can do for you next time!