BRN Discussion Ongoing

[Diogenese]

why 0, when will 1 + 1 come together?

Hi Sirod,

The Akida family are all compatible with ARM's M85, but they are not part of M85. They are optional extras.

Helium is basically ARM in-house software for machine learning and digital signal processing.

https://www.arm.com/products/silicon-ip-cpu/cortex-m/cortex-m85

Arm Cortex-M85 is the highest performing Cortex-M processor with Arm Helium technology and provides the natural upgrade path for Cortex-M based applications that require significantly higher performance and increased security.

Renesas has said that they will use their in-house DRP-AI for "heavy" AI loads, and Akida will be used for simple AI tasks. It would be surprising if Renesas were to produce the high performance M85 with only 2 nodes of Akida.

Brainchip has said that they are unable to say when the Renesas processor incorporating Akida will be produced. That is entirely within the control of Renesas.

 

[Fullmoonfever]

So, have we got an offshoot, rebranding, part the restructure process on NVISO?

BeEmotion.ai.

I thought the layout, images and graphics looked similar when I looked at their homepage initially and digging deeper find some of what I snipped below.

When I scrolled to the bottom of the page I can see the site managed by Tropus here in Perth. I know exactly where their office is as drive past it for certain client meetings.

There is also a small Japanese link at the bottom which takes you to the Japanese site.

BeEmotion

beemotion.ai

BeEmotion

beemotion.ai

HUMAN BEHAVIOUR AI​

NEUROMORPHIC COMPUTING​

BeEmotion empowers system integrators to build AI-driven human machine interfaces to transform our lives using neuromorphic computing. Understand people and their behavior in real-time without an internet connection to make autonomous devices safe, secure, and personalized for humans.

NEUROMORPHIC COMPUTING INTEROPERABILITY​

ULTRA-LOW LATENCY WITH LOW POWER​

ULTRA-LOW LATENCY (<1MS)​

Total BeEmotion Neuro Model latency is similar for GPU and BrainChip Akida™ neuromorphic processor (300 MHz), however CPU latency is approximately 2.4x slower. All models on all platforms can achieve <10ms latency and the best model can achieve 0.6ms which is almost 2x times faster than a GPU. On a clock frequency normalization basis, this represents an acceleration of 6x.

HIGH THROUGHPUT (>1000 FPS)​

BeEmotion Neuro Model performance can be accelerated by an average of 3.67x using BrainChip Akida™ neuromorphic processor at 300MHz over a single core ARM Cortex A57 as found in a NVIDIA Jetson Nano (4GB) running at close to 5x the clock frequency. On a clock frequency normalization basis, this represents an acceleration of 18.1x.

SMALL STORAGE (<1MB)​

BeEmotion Neuro Models can achieve a model storage size under 1MB targeting ultra-low power MCU system where onboard flash memory is limited. Removing the need for external flash memory saves cost and power. BrainChip Akida™ format uses 4-bit quantisation where ONNX format uses Float32 format.

DESIGNED FOR EDGE COMPUTING​

NO CLOUD REQUIRED​

PRIVACY PRESERVING​

By processing video and audio sensor data locally it does not have to be sent over a network to remote servers for processing. This improves data security and privacy as it can perform all processing disconnected from the central server, which is a more secure and private architecture decreasing security risks

 

[Frangipani]

why 0, when will 1 + 1 come together?

Let me confuse you even more:
As you will surely recall, sometimes 2 x 3 equals 4 and 3 x 3 = 6!

 

[Fullmoonfever]

So, have we got an offshoot, rebranding, part the restructure process on NVISO?

BeEmotion.ai.

I thought the layout, images and graphics looked similar when I looked at their homepage initially and digging deeper find some of what I snipped below.

When I scrolled to the bottom of the page I can see the site managed by Tropus here in Perth. I know exactly where their office is as drive past it for certain client meetings.

There is also a small Japanese link at the bottom which takes you to the Japanese site.

BeEmotion

beemotion.ai

BeEmotion

beemotion.ai

View attachment 53840

HUMAN BEHAVIOUR AI​

NEUROMORPHIC COMPUTING​

BeEmotion empowers system integrators to build AI-driven human machine interfaces to transform our lives using neuromorphic computing. Understand people and their behavior in real-time without an internet connection to make autonomous devices safe, secure, and personalized for humans.

NEUROMORPHIC COMPUTING INTEROPERABILITY​

ULTRA-LOW LATENCY WITH LOW POWER​

ULTRA-LOW LATENCY (<1MS)​

Total BeEmotion Neuro Model latency is similar for GPU and BrainChip Akida™ neuromorphic processor (300 MHz), however CPU latency is approximately 2.4x slower. All models on all platforms can achieve <10ms latency and the best model can achieve 0.6ms which is almost 2x times faster than a GPU. On a clock frequency normalization basis, this represents an acceleration of 6x.

HIGH THROUGHPUT (>1000 FPS)​

BeEmotion Neuro Model performance can be accelerated by an average of 3.67x using BrainChip Akida™ neuromorphic processor at 300MHz over a single core ARM Cortex A57 as found in a NVIDIA Jetson Nano (4GB) running at close to 5x the clock frequency. On a clock frequency normalization basis, this represents an acceleration of 18.1x.

SMALL STORAGE (<1MB)​

BeEmotion Neuro Models can achieve a model storage size under 1MB targeting ultra-low power MCU system where onboard flash memory is limited. Removing the need for external flash memory saves cost and power. BrainChip Akida™ format uses 4-bit quantisation where ONNX format uses Float32 format.

DESIGNED FOR EDGE COMPUTING​

NO CLOUD REQUIRED​

PRIVACY PRESERVING​

By processing video and audio sensor data locally it does not have to be sent over a network to remote servers for processing. This improves data security and privacy as it can perform all processing disconnected from the central server, which is a more secure and private architecture decreasing security risks

Nice to see ZF listed under the client / partner section like Brainchip.

Wouldn't mind a hook up with ZF.

Geez....they even have the BRN iceberg already

 

[Diogenese]

So, have we got an offshoot, rebranding, part the restructure process on NVISO?

BeEmotion.ai.

I thought the layout, images and graphics looked similar when I looked at their homepage initially and digging deeper find some of what I snipped below.

When I scrolled to the bottom of the page I can see the site managed by Tropus here in Perth. I know exactly where their office is as drive past it for certain client meetings.

There is also a small Japanese link at the bottom which takes you to the Japanese site.

BeEmotion

beemotion.ai

BeEmotion

beemotion.ai

View attachment 53840

HUMAN BEHAVIOUR AI​

NEUROMORPHIC COMPUTING​

BeEmotion empowers system integrators to build AI-driven human machine interfaces to transform our lives using neuromorphic computing. Understand people and their behavior in real-time without an internet connection to make autonomous devices safe, secure, and personalized for humans.

NEUROMORPHIC COMPUTING INTEROPERABILITY​

ULTRA-LOW LATENCY WITH LOW POWER​

ULTRA-LOW LATENCY (<1MS)​

Total BeEmotion Neuro Model latency is similar for GPU and BrainChip Akida™ neuromorphic processor (300 MHz), however CPU latency is approximately 2.4x slower. All models on all platforms can achieve <10ms latency and the best model can achieve 0.6ms which is almost 2x times faster than a GPU. On a clock frequency normalization basis, this represents an acceleration of 6x.

HIGH THROUGHPUT (>1000 FPS)​

BeEmotion Neuro Model performance can be accelerated by an average of 3.67x using BrainChip Akida™ neuromorphic processor at 300MHz over a single core ARM Cortex A57 as found in a NVIDIA Jetson Nano (4GB) running at close to 5x the clock frequency. On a clock frequency normalization basis, this represents an acceleration of 18.1x.

SMALL STORAGE (<1MB)​

BeEmotion Neuro Models can achieve a model storage size under 1MB targeting ultra-low power MCU system where onboard flash memory is limited. Removing the need for external flash memory saves cost and power. BrainChip Akida™ format uses 4-bit quantisation where ONNX format uses Float32 format.

DESIGNED FOR EDGE COMPUTING​

NO CLOUD REQUIRED​

PRIVACY PRESERVING​

By processing video and audio sensor data locally it does not have to be sent over a network to remote servers for processing. This improves data security and privacy as it can perform all processing disconnected from the central server, which is a more secure and private architecture decreasing security risks

This is a great endorsement of Akida!

... and BeEmotion will need to do a lot of brand recognition work and it looks like we will be dragged along with their coattails. They need us to show the benefits of their product.

So nViso must have lost out on the trade mark front to the security software mob. Maybe they thought using lower case with the capital V was sufficient to distinguish them, but, apart from being a feeble point of difference, it doesn't help phonetically.

 

[Fullmoonfever]

This is a great endorsement of Akida!

... and BeEmotion will need to do a lot of brand recognition work and it looks like we will be dragged along with their coattails. They need us to show the benefits of their product.

So nViso must have lost out on the trade mark front to the security software mob. Maybe they thought using lower case with the capital V was sufficient to distinguish them, but, apart from being a feeble point of difference, it doesn't help phonetically.

Good point.

Could be that re the trademark.

Happy to get the great wraps as if they are selling their Neuro SDK with the recommendation of running it on Akida, given some of those quick stats, then coattails it is hopefully.

 

[Frangipani]

So, have we got an offshoot, rebranding, part the restructure process on NVISO?

BeEmotion.ai.

I thought the layout, images and graphics looked similar when I looked at their homepage initially and digging deeper find some of what I snipped below.

When I scrolled to the bottom of the page I can see the site managed by Tropus here in Perth. I know exactly where their office is as drive past it for certain client meetings.

There is also a small Japanese link at the bottom which takes you to the Japanese site.

BeEmotion

beemotion.ai

BeEmotion

beemotion.ai

View attachment 53840

HUMAN BEHAVIOUR AI​

NEUROMORPHIC COMPUTING​

BeEmotion empowers system integrators to build AI-driven human machine interfaces to transform our lives using neuromorphic computing. Understand people and their behavior in real-time without an internet connection to make autonomous devices safe, secure, and personalized for humans.

NEUROMORPHIC COMPUTING INTEROPERABILITY​

ULTRA-LOW LATENCY WITH LOW POWER​

ULTRA-LOW LATENCY (<1MS)​

Total BeEmotion Neuro Model latency is similar for GPU and BrainChip Akida™ neuromorphic processor (300 MHz), however CPU latency is approximately 2.4x slower. All models on all platforms can achieve <10ms latency and the best model can achieve 0.6ms which is almost 2x times faster than a GPU. On a clock frequency normalization basis, this represents an acceleration of 6x.

HIGH THROUGHPUT (>1000 FPS)​

BeEmotion Neuro Model performance can be accelerated by an average of 3.67x using BrainChip Akida™ neuromorphic processor at 300MHz over a single core ARM Cortex A57 as found in a NVIDIA Jetson Nano (4GB) running at close to 5x the clock frequency. On a clock frequency normalization basis, this represents an acceleration of 18.1x.

SMALL STORAGE (<1MB)​

BeEmotion Neuro Models can achieve a model storage size under 1MB targeting ultra-low power MCU system where onboard flash memory is limited. Removing the need for external flash memory saves cost and power. BrainChip Akida™ format uses 4-bit quantisation where ONNX format uses Float32 format.

DESIGNED FOR EDGE COMPUTING​

NO CLOUD REQUIRED​

PRIVACY PRESERVING​

By processing video and audio sensor data locally it does not have to be sent over a network to remote servers for processing. This improves data security and privacy as it can perform all processing disconnected from the central server, which is a more secure and private architecture decreasing security risks

Great find, @Fullmoonfever!

The weird thing is - the old NVISO website www.nviso.ai is still active and lists the exact same job offer under “Careers” as www.beemotion.ai! Except for Head Of Product Bogdan Lazar’s email address, that is.
I think you may have discovered their rebranding before they went public with it!

Apparently, they are still headquartered at the same location, namely at EPFL (École Polytechnique Fédérale de Lausanne) Innovation Park in Switzerland, but also have offices in Tokyo, Dubai and Sydney.
(That info is missing on the new website, at least in the careers section).

The YPB ASX announcement dated July 27, 2023 said “Nviso SA was founded in 2009 in Switzerland with operations in Japan and recently, as part of its management and Board changes, has its Corporate Headquarters in Australia under NVISO Group Limited.”
So I assume they differentiate between their Lake Geneva research headquarters and their new head office in Sydney? I find it rather unusual that they don’t have a “Team” subpage on either website.

Engagement — BeEmotion

nviso.ai

Engagement — BeEmotion

beemotion.ai

 

[Frangipani]

This is the research institute they are or will be collaborating with:

Idiap Research Institute

The Institute is among the leaders in artificial and cognitive intelligence since 1991.

www.idiap.ch

 

[Frangipani]

And this is whom the prospective R&D Assistant will be working with…

 

[Frangipani]

So nViso must have lost out on the trade mark front to the security software mob. Maybe they thought using lower case with the capital V was sufficient to distinguish them, but, apart from being a feeble point of difference, it doesn't help phonetically.

They say “Once bitten, twice shy”.
However, the nViso / NVISO / BeEmotion management doesn’t seem to be fazed by the prospect of getting bitten again.…
Or will they possibly bee stung this time round?

 

[Sirod69]

Let me confuse you even more:
As you will surely recall, sometimes 2 x 3 equals 4 and 3 x 3 = 6!

 

[Sirod69]

AI boom fuels chip sales: What to expect in 2024​

Freund points to the emerging market players, “Don’t forget about Cerebras, or Microsoft Azure, or the Google TPU, or the AWS proprietary AI chips, Trainium and Inferentia. Or the dozen or so AI Startups beginning to ship new products in 2024, such as Groq, Samba Nova, D-Matrix, NeuReality, Untether.AI, Tenstorrent, and BrainChip.”

AI boom fuels chip sales: What to expect in 2024 - Astute Group

"Welcome to 2024: it promises to be quite the AI ride," declares Forbes contributor Karl Freund, highlighting the "explosive growth" of the AI market. This growth is fuelled by a diverse range of applications, from autonomous vehicles and edge computing to healthcare diagnostics and creative...

www.astutegroup.com

 

[MDhere]

Hi Sirod,

The Akida family are all compatible with ARM's M85, but they are not part of M85. Thay are optional extras.

Helium is basically ARM in-house software for machine learning and digital signal processing.

https://www.arm.com/products/silicon-ip-cpu/cortex-m/cortex-m85

Arm Cortex-M85 is the highest performing Cortex-M processor with Arm Helium technology and provides the natural upgrade path for Cortex-M based applications that require significantly higher performance and increased security.

Renesas has said that they will use their in-house DRP-AI for "heavy" AI loads, and Akida will be used for simple AI tasks. It would be surprising if Renesas were to produce the high performance M85 with only 2 nodes of Akida.

Brainchip has said that they are unable to say when the Renesas processor incorporating Akida will be produced. That is entirely within the control of Renesas.

Yes but Renesas has an ip licence so the can factor Akida in as they already have akida and akida is compatible with m85. So 1 +1= $
Imo

 

[Mt09]

#eyetracking #ces2024 #neuromorphic | iniVation

SynSense and iniVation announce strategic partnership for next-gen intelligent eye tracking. SynSense the leader in ultra-low-power neuromorphic processing, has teamed up with iniVation, the neuromorphic vision market leader, to build next-generation low-power, high speed eye tracking devices...

www.linkedin.com

inivation working with Synsense. CES 24 is looking a bit sad for us, nothing about Merc, Renesas or Prophesee . Hopefully our new head of sales can make some inroads..

 

[Sirod69]

I know MBUX Virtual Assistant is based on its own operating system MB.OS, but does this exclude Akida entirely?

Mercedes-Benz Media

Exklusive Einblicke und individuelle Angebote: Erleben Sie mit Mercedes-Benz das Maximum aus digitaler Live-PR. Exclusive insights and individual offers: Experience the maximum of digital live PR with Mercedes-Benz.

media.mercedes-benz.com

MBUX Virtual Assistant makes life easier with natural interaction and proactive intelligence
The groundbreaking MBUX Virtual Assistant based on MB.OS is currently the most human-like interface with a Mercedes-Benz. It connects all of MBUX's intelligent systems into a single unit. In doing so, he presents customers with a new design with natural and empathetic interaction. The MBUX Virtual Assistant is a further development of the system first introduced in VISION EQXX. It uses generative AI and proactive intelligence to make life as easy, pleasant and comfortable as possible. This allows him to make helpful suggestions based on learned behavior and situational context. This could be, for example, playing the latest news in the morning or starting a favorite massage program at the end of the work day. And if the driver is late for an appointment entered in the calendar, the system can offer to dial in directly from the vehicle. If desired, customers can also automate these processes. The experience can be further enhanced by incorporating ambient lighting and the sound system. For example, warm light and soft music can accompany a relaxing massage.

 

[Esq.111]

Morning Chippers,

Hewlett Packards latest acquisition.

https://www.reuters.com/markets/deals/hewlett-packard-enterprise-nears-13-bln-deal-buy-juniper-networks-wsj-2024-01-08/

Regards ,
Esq.

 

[Bravo]

#eyetracking #ces2024 #neuromorphic | iniVation

SynSense and iniVation announce strategic partnership for next-gen intelligent eye tracking. SynSense the leader in ultra-low-power neuromorphic processing, has teamed up with iniVation, the neuromorphic vision market leader, to build next-generation low-power, high speed eye tracking devices...

www.linkedin.com

inivation working with Synsense. CES 24 is looking a bit sad for us, nothing about Merc, Renesas or Prophesee . Hopefully our new head of sales can make some inroads..

 

[Bravo]

10 Jan 2024
Markus Schaefer, Mercedes-Benz Group CTO, discusses the incorporation of generative AI into the company's vehicles. He speaks with Ed Ludlow on "Bloomberg Technology."



2 mins "Within this system we have partners. So we're using partners for AI modules".

 

[Sirod69]

View attachment 53858

Bravo, you should rate with Hope Akida Inside , as @RegBet wrote, what's the point of this roomspray, our men can't do anything with it

 

[Bravo]

Mercedes-Benz’s best-in-class voice assistant is getting an AI boost

The company is relying on Nvidia chips to power it.

www.theverge.com

Extract only