BRN Discussion Ongoing

[IloveLamp]

Hi rgupta
If we go back to last year when some details about the Qualcomm Mercedes engagement was announced I suggested that it did not in anyway justify the belief that Mercedes had abandoned AKIDA in favour of Snapdragon.

I agree it is completely illogical to believe that Qualcomm does not have a deep understanding of Brainchip's AKIDA technology. At the very least the mutual engagement with Mercedes Benz would have force feed that understanding to Qualcomm.

When this debate concerning the merits of AKIDA v Snapdragon is had here and at HC it has by far been a debate about the merits of AKD1000 and the Qualcomm claims regarding Snapdragon's latest version/s. The reality in assessing Qualcomm's merit as a competitor in my opinion should be with respect to AKIDA 2.0 because this is Brainchip's current leading technology and Snapdragon's latest version/s are Qualcomm's leading technology in this space.

When you compare these two technologies Snapdragon still remains five years behind.

By the way loyal Peter van der Made supporters of which I am one do not think that I am in anyway disparaging the AKD1000 achievement. Remember Peter van der Made had/has a vision and it is AGI by around 2030 and Sean Hehir CEO has a plan where every 12 to 18 months a further technology step towards that destination will be commercialised. AKD1000 will always have an important historical place in the development of Brainchip the company.

To be clear my first post was a reply to @charles2 's post requesting clarification with regards to Prophesee and Branchip. I gave him in answer what I believe are indisputable propositions based upon actual knowledge of the state of play. Nothing more nothing less.

If you want speculation then my speculation is that Qualcomm just like NASA, DARPA, ARM, SiFive, Intel and Tata cannot afford to neglect consideration of AKIDA 2.0 with TENNS and ViT.

My opinion only DYOR
Fact Finder

 

[Diogenese]

We are partners with Prophesee. Any chance AKIDA is involved with the new Qualcomm Snapdragon or maybe the one after.
Prophessee are supplying a chip for Qualcomm. " Vision firm Prophesee has announced a new collaboration that will see its neuromorphic Metavision sensors optimised for use with Qualcomm’s Snapdragon mobile platforms."
BRN Partnership June14th 2022:
" “We’ve successfully ported the data from Prophesee’s neuromorphic-based camera sensor to process inference on Akida with impressive performance,” said Anil Mankar, Co-Founder and CDO of BrainChip. “This combination of intelligent vision sensors with Akida’s ability to process data with unparalleled efficiency, precision and economy of energy at the point of acquisition truly advances state-of-the-art AI enablement and offers manufacturers a ready-to-implement solution.”
Is AKIDA in the new Snapdragon described in your post?
At a presentation last year by Sean last year when talking about Tech partnerships and how we need to show our AKIDA works well with any processors etc our clients use (that is why we embedded with ARM) he said:
" a customer buys a Prophesee camera they want to know Branchip works well with them"
Not sure whether this was just a qiup or an unintentional slip of the tongue given NDA??
In any case Prophesee have been working with the AKIDA chip for almost 2 years - and it works.
I am not sure of any other event based/power saving Neuromorphic AI companies are advanced enough to have gone this far.
According to Sean at the recent presentation we are streets ahead of the competition.

Akida is capable of image classification. The deblurring feature of Prophesee/Qualcomm/Sony does not require classification.

This Prophesee patent application described merging DVS with frame camera image. Its's about time shifting the frames to eliminate blur.

Prophesee DVS Frame Camera

US2022329771A1 METHOD OF PIXEL-BY-PIXEL REGISTRATION OF AN EVENT CAMERA TO A FRAME CAMERA 20210402

A method for registering pixels provided in a pixel event stream comprising: acquiring image frames from a frame-based camera, each image frame being generated using an exposure period; generating a first point matrix from one or more of the image frames, the first point matrix being associated with an acquisition period of the image frames; acquiring a pixel event stream generated during the acquisition period; generating a second point matrix from pixel events of the pixel event stream, occurring during the acquisition period of the first point matrix; computing a correlation scoring function applied to at least a part of the points of the first and second point matrices, and estimating respective positions of points of the second point matrix in the first point matrix, due to depths of the points of the first point matrix related to the second point matrix, by maximizing the correlation scoring function.

 

[Diogenese]

I have posed the same question in the past @Bravo

Perhaps it is part of an NDA that BRN signed that prohibits them from mentioning Qualcomm in any form? Just a thought

I for one think the fact Qualcomm has never been mentioned as a competitor, even though their recent claims suggest they should be, leaves me with a most comfortable feeling.....


View attachment 58155

My opinion only. Dyor

EDIT: P.S

The same could be said for APPLE and SAMSUNG...........

Hi ILL,

That comforting feeling drops away as the temperature drops.

 

[IloveLamp]

Hi ILL,

That comforting feeling drops away as the temperature drops.

 

[Esq.111]

Afternoon Diogenese ,

Thankyou once again for your tireless Technical Knowledge , mind boggling.

From what i have picked up .... and i could be mistaken.... our chip / IP works on First Kernel in is First Kernel out , with No Backend Loading ( whatever that means ).

As per picture ......

Regards,
Esq.

 

[IloveLamp]

Green..........

 

[7für7]

Actually it looks like Competitor 1 is Intel, if the previous chart is anything to go by. How come Qualcomm has never been on any of our competitor charts?


View attachment 58152

What the hack are you doing here!? No wonder the share price drops please do us a favour oookeeeeeyyy?!?!?!

 

[IloveLamp]

On-device AI | Technologies | Samsung Semiconductor Global

Discover the power of on-device AI technology, enabling seamless processing directly on your mobile device. Visit Samsung Semiconductor to learn more.

semiconductor.samsung.com

Exynos2400: On-device AI | Samsung Semiconductor

Generate images from text, express yourself more clearly with the help of AI, and even turn the color of a hat in a photo from 🟩 to 🟪, all without data leaving your device. Learn more about on-device AI with the #Exynos2400 from #SamsungSemiconductor. https://lnkd.in/gRNSicGJ

www.linkedin.com

 

[Diogenese]

Prophesee has their own Neuromorphic ip, although i believe ours is superior hense why they partnered with us.......


So it is possible it excludes us without including anyone else........but i like our chances

Prophesee has also never been mentioned by the company as a competitor either..........

That is my understanding

Qualcomm has their own AI, Hexagon, in Snapdragon 8.2.


Snapdragon 8 Gen 2 deep dive: Everything you need to know (androidauthority.com)

Snapdragon 8 Gen 2 deep dive: Everything you need to know

All you'll ever need to know about Qualcomm's latest Snapdragon 8 Gen 2 mobile platform for next-gen phones can be found here.

www.androidauthority.com

To boost performance, the Tensor accelerator inside the DSP has doubled in size for twice the performance and has new optimizations specifically for language processing. Qualcomm is also debuting what it calls micro tile inference support, essentially chopping up imaging and other problems into smaller tiles to save on memory at the expense of some result accuracy. Along those lines, the addition of INT4 also means that developers can now implement machine learning problems requiring high bandwidth at the expense of some accuracy if compressing a larger model. Qualcomm is providing tools to partners to help support INT4, so it will require a retooling of existing applications to work.

Altogether, the Snapdragon 8 Gen 2 Hexagon DSP offers 4.35x the performance of its predecessor, depending on the ML model. (In this case, Qualcomm is comparing mobileBERT natural language processing). Sounds impressive, but I think the more significant change is the introduction of Hexagon Direct Link, which more closely connects its ISP to the AI Engine. The company dubs this its “Cognitive ISP.”

...
Qualcomm doubled the physical link between the image signal processor (ISP), Hexagon DSP, and Adreno GPU, driving higher bandwidth and lowering latency. This allows the Snapdragon 8 Gen 2 to run much more powerful machine-learning tasks on imaging data right off the camera sensor. RAW data, for instance, can be passed directly to the DSP/AI Engine for imaging workloads, or Qualcomm can use the link to upscale low-res gaming scenarios to assist with GPU load balancing.

US2020104690A1 NEURAL PROCESSING UNIT (NPU) DIRECT MEMORY ACCESS (NDMA) HARDWARE PRE-PROCESSING AND POST-PROCESSING 20180928

[0015] FIG. 4 is a block diagram illustrating an exemplary software architecture that may modularize artificial intelligence (AI) functions, in accordance with aspects of the present disclosure.
...
In the exemplary example, the deep neural network may be configured to run on a combination of processing blocks, such as the CPU 422 , the DSP 424 , and the GPU 426 , or may be run on the NPU 428 .
...
. In aspects of the present disclosure, the read client RCLT and the write client WCLT may refer to an array to compute elements of the NPU 700 , which may support, for example, 16-NDMA read channels and 16-NDMA write channels for the various compute units of the NPU 700 .

[0021] FIG. 7 is a block diagram illustrating a neural processing unit (NPU), including an NPU direct memory access (NDMA) core and interfaces configured to provide hardware pre-processing and post-processing of NDMA data, according to aspects of the present disclosure.

 

[Diogenese]

View attachment 58161


View attachment 58162

I got the mushrooms from a little shop in Chinatown and I've lost the receipt.

 

[Boab]

I got the mushrooms from a little shop in Chinatown and I've lost the receipt.

Magic

 

[Fact Finder]

My wild theory for the day.

What if Ericsson, Brainchip and Prophesee were working together to substitute Prophesee’s event based neuromorphic vision sensor for the camera used in the above paper.

My technophobe understanding is this would further reduce power consumption making it a World beating best fit for being powered by energy harvesting from normal internal light sources.

My opinion only DYOR
Fact Finder

Hi All
My wild theory just became a little less wild.

In 2022 these Ericsson researchers published a paper which according to the Abstract was exploring:

Resilience Analysis of Distributed Wireless Spiking Neural Networks

Spiking neural networks (SNN) are expected to enable several use-cases in future communication networks (beyond 5G and 6G), as edge AI and battery-constrained systems can leverage the fast computation and high-power efficiency offered by SNNs. In this work we consider a Distributed Wireless SNN...

ieeexplore.ieee.org

Tamás Borsos
Ericsson Research

Massimo Condoluci
Ericsson Research

Marios Daoutis
Ericsson Research

Péter Hága
Ericsson Research

András Veres
Ericsson Research

“Abstract:Spiking neural networks (SNN) are expected to enable several use-cases in future communication networks (beyond 5G and 6G), as edge AI and battery-constrained systems can leverage the fast computation and high-power efficiency offered by SNNs. In this work we consider a Distributed Wireless SNN (DW-SNN) system and we analyze its performance in terms of inference accuracy and total neural activity when radio losses are applied to spikes transferred during the inference phase. Our aim is to understand how radio losses impact performance when considering different SNN spike communication types, i.e., input, excitatory, and inhibitory spikes. Then we evaluate the impact of different traffic prioritization approaches among SNN spikes when considering a shared channel capacity being available for SNN activity. From these analyses, we derive some key insights and features that can be considered when applying a DW-SNN and handling its traffic over wireless communication systems. Finally, we report a prototype implementation of DW-SNN using custom-built IoT components, which we use to further investigate different coverage scenarios.View less”

Under References they were aware of:

5.
"Prophesee Metavision for Machines", [online] Available: https://www.prophesee.ai/.
Hide Context Google Scholar
In a neuromorphic or event camera [4][5]–[6] each pixel directly feeds corresponding neuron(s) and these neurons emit spikes when the change in light intensity exceeds a predefined threshold [6].
Go To Text

Unfortunately I cannot access the full paper and the Prophesee referenced paper appears to have been removed or blocked.

My opinion only DYOR
Fact Finder

 

[Quiltman]

 

[Fact Finder]

Hi All
My wild theory just became a little less wild.

In 2022 these Ericsson researchers published a paper which according to the Abstract was exploring:

Resilience Analysis of Distributed Wireless Spiking Neural Networks

Spiking neural networks (SNN) are expected to enable several use-cases in future communication networks (beyond 5G and 6G), as edge AI and battery-constrained systems can leverage the fast computation and high-power efficiency offered by SNNs. In this work we consider a Distributed Wireless SNN...

ieeexplore.ieee.org

Tamás Borsos
Ericsson Research

Massimo Condoluci
Ericsson Research

Marios Daoutis
Ericsson Research

Péter Hága
Ericsson Research

András Veres
Ericsson Research

“Abstract:Spiking neural networks (SNN) are expected to enable several use-cases in future communication networks (beyond 5G and 6G), as edge AI and battery-constrained systems can leverage the fast computation and high-power efficiency offered by SNNs. In this work we consider a Distributed Wireless SNN (DW-SNN) system and we analyze its performance in terms of inference accuracy and total neural activity when radio losses are applied to spikes transferred during the inference phase. Our aim is to understand how radio losses impact performance when considering different SNN spike communication types, i.e., input, excitatory, and inhibitory spikes. Then we evaluate the impact of different traffic prioritization approaches among SNN spikes when considering a shared channel capacity being available for SNN activity. From these analyses, we derive some key insights and features that can be considered when applying a DW-SNN and handling its traffic over wireless communication systems. Finally, we report a prototype implementation of DW-SNN using custom-built IoT components, which we use to further investigate different coverage scenarios.View less”

Under References they were aware of:

5.
"Prophesee Metavision for Machines", [online] Available: https://www.prophesee.ai/.
Hide Context Google Scholar
In a neuromorphic or event camera [4][5]–[6] each pixel directly feeds corresponding neuron(s) and these neurons emit spikes when the change in light intensity exceeds a predefined threshold [6].
Go To Text

Unfortunately I cannot access the full paper and the Prophesee referenced paper appears to have been removed or blocked.

My opinion only DYOR
Fact Finder

Hi All
Less wild is the theory I and others have advanced that MegaChips is working with Morse Micro to add AKIDA technology into their WiFi HaLow chip offering.

This paper from Ericsson also adds weight to the merit of such speculation.

My opinion only DYOR
Fact Finder

 

[Slade]

BrainChip DJ

Pump up the BrainChip, pump it up While the stock's jumpin'
And the value's pumpin'
Look ahead, the charts are thumpin'
Pump it up a little more
Get the bulls on the floor Pump up the BrainChip, pump it up
While the market's boomin'
And the bears are gloomin'
Reach the heights, no room for losin'
Pump it up a little more Get the bulls on the floor

 

[Bravo]

I got the mushrooms from a little shop in Chinatown and I've lost the receipt.

That's so weird. I got mine from an Asian grocery store in Melbourne and I've lost my receipt too!!!

 

[Damo4]

 

[Bravo]

What the hack are you doing here!? No wonder the share price drops please do us a favour oookeeeeeyyy?!?!?!

OK, I don't need to be told twice!!! It'll only be a short run (i.e fast motion walk) around the block though because I have a painful bunion on my left foot and an extremely large corn on my right foot. Back soon!

 

[Rskiff]

OK, I don't need to be told twice!!! It'll only be a short run (i.e fast motion walk) around the block though because I have a painful bunion on my left foot and an extremely large corn on my right foot. Back soon!

Those green boots causing bunions?

 

[supersonic001]

#edgeai #ai #neuromorphic | BrainChip | 10 comments

Shh, it's a sneak peek! We visited the BrainChip office recently to get you a little behind-the-scenes glimpse of BrainChip’s engineering in action and the Akida Edge AI Box. Pre-orders are open. - https://lnkd.in/dxMtzuEX #EdgeAI #AI #neuromorphic | 10 comments on LinkedIn

www.linkedin.com