BRN Discussion Ongoing

[Fact Finder]

Hi FF,
Nintendo is a possibility in saying that what about PlayStation - Sony is Japanese!!

I read something somewhere that claimed Sony was in some way shape or form connected with MegaChips but could not find anything that I would call solid evidence. Beyond that you have your Japanese connection though it is a bit like when you travel overseas and someone finds out your are Australian living in Sydney and say do you know Bill Jones he is a mate of mine and he lives in Perth.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[thelittleshort]

These big tech players are right until they are wrong. Intel was right for a long time but now they are struggling. Facebook now trying to reinvent itself as Meta was right until it is now wrong and is loosing market share and trying to reinvent itself. These huge companies come and go and get it wrong in the end. Apple had a dark period. Far too many to mention in the tech space but Nokia springs to mind. Incumbency carries as much risk as being new and revolutionary. LeCunn is wedded to his historical position. Others have no idea because it was not taught in university when they obtained their degrees 20 or 30 years ago. Conventional wisdom is a greater inhibitor of innovation.

The facts are that Mercedes, NASA, Valeo and Vorago are early adopters and Renesas and MegaChips have licenced the IP. Nanose and Biotome are impressed enough to take Brainchip on as a strategic partner. The product is real. The product is in the market place. Brainchip is now shipping it to universities for academics to play with. Intel and IBM have been trying to do what Brainchip has achieved. Socionext, Eastronics, Saleslink have been convinced that AKIDA technology has a role to play. Not to mention Information Systems Laboratories have broken cover admitting they have been an EAP and together with Brainchip are developing a radar solution for the US Airforce under a research grant.

About a month ago I was reading an article about the presently accepted theory of our expanding Universe. The scientist and philosopher who was writing the article was actually making the point that though the present theory stacks up and he believes it if he were to be born 100 million years from now the expanding universe would have expanded to such an extent that absent current knowledge and measurements the other universes that can be presently seen with existing technology will be so far away that they will no longer exist in the sense they will be beyond the range of scientists to see them or the signals they emit.

So that scientists in 100 million years will if all present knowledge is lost will necessarily need a different theoretical construct and will come to an entirely different conclusion. The point he is making is that current scientists do not know what they do not know and thus things may have occurred in the past that should be in their models but are not and never will be.

The AKIDA technology did not exist before Peter van der Made existed. He has manufactured out of thin air the science underpinning it and why it works and that science is patent protected and covered by trade secrets. It is impossible for others like LeCunn to actually comment because he can only have flawed data. It is very simple. If I give you a phone without a sim card you cannot make a telephone call. If I give you sand with out cement you cannot make concrete. Flawed and or incomplete data does not allow you to analyse or test a scientific theory. To this very day many of Einstein's theories are still to be tested because they were too advanced for existing technology to replicate but he so far continues to be correct.

If Sir Winston Churchill and the British people had believed their competitors about their chances Germany, Italy and Japan would have won the Second World War. Competitors have a vested interest in denying the strengths of their competitors.

My opinion only DYOR
FF

AKIDA BALLISTA

Fantastic response @Fact Finder

 

[JK200SX]

There is one more question you could ask: So is the C Class still running the MBUX voice recognition system that motoring writers have described as clunky and not very accurate?”

Would be interesting to have the answer.

Regards
FF

 

[Fact Finder]

View attachment 2364
View attachment 2365

View attachment 2366
View attachment 2367
View attachment 2368
View attachment 2369
View attachment 2371
View attachment 2372
View attachment 2373
View attachment 2374

Such hard work to drag that answer out. Great work it had me laughing my head off. So they have not been completely honest in saying it is the first time implying first time across their range only first time on the C Class.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Fact Finder]

Hi Esq.111,

I did start to draft a response to your post about R_CAR V4H last night but nodded off.

I didn't see the article you referred to but I did check the Renesas website and saw, as you mentioned, it's a couple of years away.

I would think that that would be sufficient time to include Akida IP. A couple of pages ago I posted the R-CAR V4H block diagram which showed a CNN IP block. Now that may well be a way of concealing Akida CNN2SNN behind the CNN veil.

After all, who else does CNN in silicon?

Rubber maybe but silicon never I get a rash. FF

 

[Flenton]

View attachment 2364
View attachment 2365

View attachment 2366
View attachment 2367
View attachment 2368
View attachment 2369
View attachment 2371
View attachment 2372
View attachment 2373
View attachment 2374

How about the reverse psychology approach to find out if it's not in there and something like...

I heard the upgraded MBUX is running of something called Akida. My reseach tells me this is not stable can you confirm it is not using this technology?

 

[Dang Son]

20 and a half million shares later ......... Gee Whiz ...... up 2.5c !!!
What a load of manipulation but I am not suprised.

Seriously, how many of the 3099 trades (sales) were genuine ? Not many I reckon.
Patience is the key here, I realise this - but does not mean I have to like it and unfair tactics by some in any domain bothers me very much.
I have some minor suggestions for changes to the trading info on the ASX platform.
I am curious to understand if this may be doo-able without too much difficulty ....... or am asking for too much ?

1. REGISTERED NUMBERS to buy and sell shares (no personal details) ........ just like an auction and the registered number included with each individual sale/purchase.
2. SiGNAL the direction of all transactions ........ if they are in direction of the buyer price or seller price.

Would that be difficult to do ?
Would that be more fair and transparent ........... or would that be unfair and too transparent for the manipulators ?
Just wondering if this would be a reasonable proposal to put to the ASX ......... IN THE NAME OF LESS MANIPULATION ?

I presume the ASX wants to promote and endorse a fair operation of its platform and by refusing the above suggestions would be a signal
that they oppose - LESS MANIPULATION, and support MORE MANIPLUATION ..... would that be true ?

Hi newpunta,
Definitely ASX daily trading data would show that Bots are responsible for BRN manipulated capitulation of 50%.
Weather it's Shorters or Accumulators it's plain to see, it was their Bots which made their gain possible and our loss in a big way.
Over and over through the days , a 1 share sale will drop the price a pip or two , then come to after market adjust, their Bots have agility to know when 1 more share on the sell side will lower the closing price, every day practically and nothing we can do about it.
50% is like 1.5b that was just ripped off from Aussie retail because ASIC licences Bots without rules, because someone in Authority took an underhand payment for a green light, IMO
Sadly in need of a judicial Inquiry into the domestic cost of Algorithmic Trading and expose' of all Party's involved.
I just had a late thought that the govt probably makes so much tax from Bot operators at the end of the day and it cares not how much citizens lose.

 

[Fact Finder]

These big tech players are right until they are wrong. Intel was right for a long time but now they are struggling. Facebook now trying to reinvent itself as Meta was right until it is now wrong and is loosing market share and trying to reinvent itself. These huge companies come and go and get it wrong in the end. Apple had a dark period. Far too many to mention in the tech space but Nokia springs to mind. Incumbency carries as much risk as being new and revolutionary. LeCunn is wedded to his historical position. Others have no idea because it was not taught in university when they obtained their degrees 20 or 30 years ago. Conventional wisdom is a greater inhibitor of innovation.

The facts are that Mercedes, NASA, Valeo and Vorago are early adopters and Renesas and MegaChips have licenced the IP. Nanose and Biotome are impressed enough to take Brainchip on as a strategic partner. The product is real. The product is in the market place. Brainchip is now shipping it to universities for academics to play with. Intel and IBM have been trying to do what Brainchip has achieved. Socionext, Eastronics, Saleslink have been convinced that AKIDA technology has a role to play. Not to mention Information Systems Laboratories have broken cover admitting they have been an EAP and together with Brainchip are developing a radar solution for the US Airforce under a research grant.

About a month ago I was reading an article about the presently accepted theory of our expanding Universe. The scientist and philosopher who was writing the article was actually making the point that though the present theory stacks up and he believes it if he were to be born 100 million years from now the expanding universe would have expanded to such an extent that absent current knowledge and measurements the other universes that can be presently seen with existing technology will be so far away that they will no longer exist in the sense they will be beyond the range of scientists to see them or the signals they emit.

So that scientists in 100 million years will if all present knowledge is lost will necessarily need a different theoretical construct and will come to an entirely different conclusion. The point he is making is that current scientists do not know what they do not know and thus things may have occurred in the past that should be in their models but are not and never will be.

The AKIDA technology did not exist before Peter van der Made existed. He has manufactured out of thin air the science underpinning it and why it works and that science is patent protected and covered by trade secrets. It is impossible for others like LeCunn to actually comment because he can only have flawed data. It is very simple. If I give you a phone without a sim card you cannot make a telephone call. If I give you sand with out cement you cannot make concrete. Flawed and or incomplete data does not allow you to analyse or test a scientific theory. To this very day many of Einstein's theories are still to be tested because they were too advanced for existing technology to replicate but he so far continues to be correct.

If Sir Winston Churchill and the British people had believed their competitors about their chances Germany, Italy and Japan would have won the Second World War. Competitors have a vested interest in denying the strengths of their competitors.

My opinion only DYOR
FF

AKIDA BALLISTA

The problem with these ageing engineers at Nvidia and other places is they have not yet learned how to do a Google search. I being a spritely 68 year old retired lawyer however and totally switched on to the world of technology have mastered a simple Google search and in a wink of an eye found the following paper from IMEC proving that an SNN using 4 bit can out perform DNN at the extreme edge analysing radar signals for hand gesture.

VII. CONCLUSION This article has shown the impact of various preprocessing and event-encoding techniques on the accuracy of a 4-b-weight SNN within the context of radar gesture recognition for extreme-edge applications with limited power and die area. In particular, it has been shown that, while using the same network topology and learning approach, preprocessing drastically affects the SNN accuracy, while its effect is significantly more limited on the DNN accuracy. In addition, it has been demonstrated that conclusions drawn from DNN design cannot directly be generalized to SNN design as the preprocessing parameters can have antagonistic effects on the SNN and DNN accuracy, even if the same topology and learning approaches are used. Also, the impact of event-encoding to optimize the accuracy of the network has been explored, be it that it has less impact compared to preprocessing. Then, after extensive comparison between different approaches, a well-suited radar preprocessing and the event-encoding queue has been developed to be used with a small, 4-b-weight SNN to achieve 93% of recognition accuracy within only four time steps. It must be noted that the preprocessing and event-encoding strategy developed in this article could also be extended to audio signals as future work. Finally, the applicability of the proposed Authorized licensed use limited to:

IMEC. Downloaded on September 16,2021 at 17:52:11 UTC from IEEE Xplore. Restrictions apply. This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

12 IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 4-b-weight SNN approach has been demonstrated on a second radar dataset, reaching 94.6% of accuracy. Unlike previous works, this low-complexity architecture enables high-accuracy and low-latency inference while consuming very little energy and area when implemented in event-based hardware, making it extremely useful for embedded edge-AI applications.

https://scholar.google.com.au/scholar_url?url=https://www.researchgate.net/profile/Ali-Safa-5/publication/354588752_Improving_the_Accuracy_of_Spiking_Neural_Networks_for_Radar_Gesture_Recognition_Through_Preprocessing/links/614384ffa609b152aa156859/Improving-the-Accuracy-of-Spiking-Neural-Networks-for-Radar-Gesture-Recognition-Through-Preprocessing.pdf&hl=en&sa=X&ei=VE4oYpDnKtCsywSO1YzoDg&scisig=AAGBfm0ntzNeCluTdKaEsSkxPUQejoRasA&oi=scholarr

I am not sure about the date of this article but it is after 2020 but before 2021 which causes me to wonder if ISL was influenced by these findings to decide to come on board with Brainchip as an EAP for creating the COTS AKD1000 Radar solution for the US Airforce.

Somebody apart from me must have read it even if the people at Nvidia cannot master Google searches.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Fact Finder]

The problem with these ageing engineers at Nvidia and other places is they have not yet learned how to do a Google search. I being a spritely 68 year old retired lawyer however and totally switched on to the world of technology have mastered a simple Google search and in a wink of an eye found the following paper from IMEC proving that an SNN using 4 bit can out perform DNN at the extreme edge analysing radar signals for hand gesture.

VII. CONCLUSION This article has shown the impact of various preprocessing and event-encoding techniques on the accuracy of a 4-b-weight SNN within the context of radar gesture recognition for extreme-edge applications with limited power and die area. In particular, it has been shown that, while using the same network topology and learning approach, preprocessing drastically affects the SNN accuracy, while its effect is significantly more limited on the DNN accuracy. In addition, it has been demonstrated that conclusions drawn from DNN design cannot directly be generalized to SNN design as the preprocessing parameters can have antagonistic effects on the SNN and DNN accuracy, even if the same topology and learning approaches are used. Also, the impact of event-encoding to optimize the accuracy of the network has been explored, be it that it has less impact compared to preprocessing. Then, after extensive comparison between different approaches, a well-suited radar preprocessing and the event-encoding queue has been developed to be used with a small, 4-b-weight SNN to achieve 93% of recognition accuracy within only four time steps. It must be noted that the preprocessing and event-encoding strategy developed in this article could also be extended to audio signals as future work. Finally, the applicability of the proposed Authorized licensed use limited to:

IMEC. Downloaded on September 16,2021 at 17:52:11 UTC from IEEE Xplore. Restrictions apply. This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

12 IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 4-b-weight SNN approach has been demonstrated on a second radar dataset, reaching 94.6% of accuracy. Unlike previous works, this low-complexity architecture enables high-accuracy and low-latency inference while consuming very little energy and area when implemented in event-based hardware, making it extremely useful for embedded edge-AI applications.

https://scholar.google.com.au/scholar_url?url=https://www.researchgate.net/profile/Ali-Safa-5/publication/354588752_Improving_the_Accuracy_of_Spiking_Neural_Networks_for_Radar_Gesture_Recognition_Through_Preprocessing/links/614384ffa609b152aa156859/Improving-the-Accuracy-of-Spiking-Neural-Networks-for-Radar-Gesture-Recognition-Through-Preprocessing.pdf&hl=en&sa=X&ei=VE4oYpDnKtCsywSO1YzoDg&scisig=AAGBfm0ntzNeCluTdKaEsSkxPUQejoRasA&oi=scholarr

I am not sure about the date of this article but it is after 2020 but before 2021 which causes me to wonder if ISL was influenced by these findings to decide to come on board with Brainchip as an EAP for creating the COTS AKD1000 Radar solution for the US Airforce.

Somebody apart from me must have read it even if the people at Nvidia cannot master Google searches.

My opinion only DYOR
FF

AKIDA BALLISTA

Sorry I just cannot help putting the boot into those with closed minds and vested interest that sees them sacrifice integrity for ignorance and self interest. Here is another example of SNN beating the crap out of conventional computing this time being proven by TATA Consulting Services Robotics Research Team:

Abstract:

In the past two decades, radar-based human sensing has become a topic of intense research. Unlike vision-based techniques which require the use of camera, radars are unobtrusive and privacy preserving in nature. Further, radars are agnostic of the lighting conditions and can be used for through-the-wall imaging thereby making them hugely effective in many situations. Compact, affordable radars have been designed that can be easily integrated with remote monitoring systems. However, the classical machine learning techniques currently used for learning and inferring human actions from radar images are compute intensive, and require large volume of training data, making them unsuitable for deployment on the network edge. In this paper, we propose to use the concepts of neuromorphic computing and Spiking Neural Networks (SNN) to learn human actions from data captured by the radar. To the best our knowledge, this is the first attempt of using SNNs on micro-Doppler data from radars. Our SNN model is capable of learning spatial as well as temporal features from the data and our experiments have resulted in 85% accuracy which is comparable with the classical machine learning approaches that are typically used on similar data. Further, the use of neuromorphic and SNN concepts make our model deployable over evolving neuromorphic edge devices thereby making the entire approach more efficient in terms of data, computation and energy consumption.

https://scholar.google.com.au/scholar_url?url=https://ieeexplore.ieee.org/abstract/document/9206853/&hl=en&sa=X&ei=VE4oYpDnKtCsywSO1YzoDg&scisig=AAGBfm165r9sqOe9LzyH75QZzw5vvlLAaQ&oi=scholarr

My opinion only but maybe I am being too harsh perhaps where these Nvidia guys work they do not have the internet yet.
FF

AKIDA BALLISTA

 

[thelittleshort]

In case you have not yet seen the other companies represented by Eastronics under:

Embedded Solutions - Components & Embedded Systems

Some heavy hitters in there…with one more recently added

 

[Slade]

In case you have not yet seen the other companies represented by Eastronics under:

Embedded Solutions - Components & Embedded Systems

Some heavy hitters in there…with one more recently added

View attachment 2375

And this is how they describe Brainchip on the Eastronics site:

 

[newpunta]

Hi newpunta,
Definitely ASX daily trading data would show that Bots are responsible for BRN manipulated capitulation of 50%.
Weather it's Shorters or Accumulators it's plain to see, it was their Bots which made their gain possible and our loss in a big way.
Over and over through the days , a 1 share sale will drop the price a pip or two , then come to after market adjust, their Bots have agility to know when 1 more share on the sell side will lower the closing price, every day practically and nothing we can do about it.
50% is like 1.5b that was just ripped off from Aussie retail because ASIC licences Bots without rules, because someone in Authority took an underhand payment for a green light, IMO
Sadly in need of a judicial Inquiry into the domestic cost of Algorithmic Trading and expose' of all Party's involved.
I just had a late thought that the govt probably makes so much tax from Bot operators at the end of the day and it cares not how much citizens lose.

Thanks Dang,

I think I understand in a small way.......... especially the bit about good for govt revenue but not for citizens.
Bots will be Bots I guess ......... didn't realise we also have our own to push things up.

Bots seem to me to only be very small volumes, even though they will all add up to a tidy sum by days end .......... the majority of the volume will come from us humans.
Humans and for that matter bots - with an ASX registered number showing next to the trade would help create a more fair view of trading patterns.
Transparency with Anonymity ......... Government still gets the tax.
Just my layman idea ..... to help not stop the rorting/shorting, but to possibly show it.

 

[Baisyet]

These big tech players are right until they are wrong. Intel was right for a long time but now they are struggling. Facebook now trying to reinvent itself as Meta was right until it is now wrong and is loosing market share and trying to reinvent itself. These huge companies come and go and get it wrong in the end. Apple had a dark period. Far too many to mention in the tech space but Nokia springs to mind. Incumbency carries as much risk as being new and revolutionary. LeCunn is wedded to his historical position. Others have no idea because it was not taught in university when they obtained their degrees 20 or 30 years ago. Conventional wisdom is a greater inhibitor of innovation.

The facts are that Mercedes, NASA, Valeo and Vorago are early adopters and Renesas and MegaChips have licenced the IP. Nanose and Biotome are impressed enough to take Brainchip on as a strategic partner. The product is real. The product is in the market place. Brainchip is now shipping it to universities for academics to play with. Intel and IBM have been trying to do what Brainchip has achieved. Socionext, Eastronics, Saleslink have been convinced that AKIDA technology has a role to play. Not to mention Information Systems Laboratories have broken cover admitting they have been an EAP and together with Brainchip are developing a radar solution for the US Airforce under a research grant.

About a month ago I was reading an article about the presently accepted theory of our expanding Universe. The scientist and philosopher who was writing the article was actually making the point that though the present theory stacks up and he believes it if he were to be born 100 million years from now the expanding universe would have expanded to such an extent that absent current knowledge and measurements the other universes that can be presently seen with existing technology will be so far away that they will no longer exist in the sense they will be beyond the range of scientists to see them or the signals they emit.

So that scientists in 100 million years will if all present knowledge is lost will necessarily need a different theoretical construct and will come to an entirely different conclusion. The point he is making is that current scientists do not know what they do not know and thus things may have occurred in the past that should be in their models but are not and never will be.

The AKIDA technology did not exist before Peter van der Made existed. He has manufactured out of thin air the science underpinning it and why it works and that science is patent protected and covered by trade secrets. It is impossible for others like LeCunn to actually comment because he can only have flawed data. It is very simple. If I give you a phone without a sim card you cannot make a telephone call. If I give you sand with out cement you cannot make concrete. Flawed and or incomplete data does not allow you to analyse or test a scientific theory. To this very day many of Einstein's theories are still to be tested because they were too advanced for existing technology to replicate but he so far continues to be correct.

If Sir Winston Churchill and the British people had believed their competitors about their chances Germany, Italy and Japan would have won the Second World War. Competitors have a vested interest in denying the strengths of their competitors.

My opinion only DYOR
FF

AKIDA BALLISTA

And what springs to mind as Eastman Kodak, they actually developed Digital camera Engeneer was Steven Sasson. Kodak missed a trick by not developing this technological breakthrough, as it chose to continue to focus on photographic film. And rest is history. Like you said FF early adopters are the winners if the Tech has been varified and works. Once you missed the opportunity you miss the whole thing.
Cheers

 

[RobjHunt]

Why are these things not released to market!!!? I’m baffled

Great stuff and good times ahead for Brainchip, I just don’t understand how this is not an ASX announcement????

Ps. Well done to whoever found this. I saw it first on HC posted by no other than Sunny who hs completely changed his tune again. Lol

I’m sorry but I’ve missed what you are commenting about??

 

[Diogenese]

Afternoon Diogenese,

Sounds like you might have to grab him by the ears, drag him outside and give him a jolly flogging with your club till he sees sense

Have you seen the EENews article from yesterday. Renesas R-Car V4H ?

Regards,
Esq

Re: R-Car V4H

Just to add to my previous post, for comparison, here are the V3H and V4H block diagrams:

V3H has a CNN hardware accelerator

V4H has CNN-IP with 2 MB memory.

Has anyone seen anybody offering CNN-IP hardware licences?

PS: You can limit your search if the pre-condition of that "anybody" is:
having a licence agreement with Renesas.

 

[Iseki]

I am hoping the 1000 eyes can take another glance at the Megachips deal.

Yes 87% of Megachips income is via Nintendo, but their webpage has a lot more to say about new routers that will connect all of the internet, the IoT, devices that mesh together etc. Router protocols and their algorithms shunt the data hoping they have managed to find a short, fast connection. New protocols are being designed to be adaptive to the many paths, and sudden changes in connectivity.

Might SNN's be able to solve these types of problems in a dynamic, adaptive, self learning, way?

I can find some academic references such as

Neural network based optimal routing algorithm for communication networks​

Neural network based optimal routing algorithm for communication networks

This paper presents the capability of the neural networks as a computational tool for solving constrained optimization problem, arising in routing alg…

www.sciencedirect.com

But sadly it's beyond me.

I'm hoping that there is a clear line of sight between BRN's deal with Megachips and millions new routers needed to cope with the foreseen increase in internet traffic. Can anyone see it?

 

[TheFunkMachine]

I’m sorry but I’ve missed what you are commenting about??

 

[Frank Zappa]

Hi FF,
Thank you for your underpants analogy, for some of us it is a huge problem.
BRN best stay out of Scotland though! I am reliably informed the Scots wear nothing under them kilts.
VG

That is no longer the case at my favourite Scottish Restaurant

 

[Frank Zappa]

That is no longer the case at my favourite Scottish Restaurant

MacDonalds

 

[Esq.111]

Re: R-Car V4H

Just to add to my previous post, for comparison, here are the V3H and V4H block diagrams:
View attachment 2379

View attachment 2380
V3H has a CNN hardware accelerator

V4H has CNN-IP with 2 MB memory.

Has anyone seen anybody offering CNN-IP hardware licences?

PS: You can limit your search if the pre-condition of that "anybody" is:
having a licence agreement with Renesas.

Evening Diogenese,

Cheers for that , I'm a very slow learner when it comes to computer tech though I do enjoy the pictures & the flow on applications of such.

Also the knowledge of those that know what the hell thay are looking at.

Regards,
Esq.