BRN Discussion Ongoing

[MDhere]

My apparent vigil was simply due to the fact that I live in a different time zone from you…
So it was actually way before my usual bedtime when I hit the send button.
Saying that, I have indeed posted at ungodly hours before, after getting myself into a state of sleuthing flow and totally losing track of time. Luckily, I get by on very little sleep…

I hope that I still get to keep the award, though?

Come to Australia at the next Agm and i will have the award ready for u i will mark it in my diary

 

[MDhere]

Hmmm taking about diarys I read somewhere can't remember where or what it about but i believe brn might need me to come along as an interpreter in Biarritz in March 2024 Don't mind me ive come back from another cold melb trip to sunny 28 degrees brissie weather with some pinot gris basking in the sun preparing for my Biarritz tax deduction business trip. is that permitted as brn is an invesrnent and im going for research ??

 

[Pappagallo]

Are BrainChip shares finally cheap enough to buy?

BrainChip Holdings Ltd (ASX: BRN) shares have become cheaper and cheaper throughout 2023. Is it finally time to buy this AI stock?

www.fool.com.au

Claim “uncertainty” about Akida 2.0 but don’t elaborate, just reference a previous MF article that also claims “uncertainty” but again doesn’t elaborate.

Lol these MF articles really are dumb as dogshit.

 

[JB49]

https://www.diva-portal.org/smash/get/diva2:1779206/FULLTEXT01.pdf

Not sure if shared before. Decent study, but noted that its still early days. Suggested using other architectures that could perhaps yield more promising results than akida.

 

[Iseki]

Claim “uncertainty” about Akida 2.0 but don’t elaborate, just reference a previous MF article that also claims “uncertainty” but again doesn’t elaborate.

Lol these MF articles really are dumb as dogshit.

Well Brainchip tapes out akd1500 chip means that the chip has been through a lot of theoretical tests and it has passed. ie there isn't much risk about actually making them. But there is that last step. Brainchip could spend $4mill and produce a pile of them and hand them out. So MF are casting shade on Brainchip for not wasting $4mill in this exercise which isn't necessary if Brainchip know there are already other OEM's going to manufacture the chip.

 

[manny100]

Great yes but this should be looked from inside to out imo; if it’s Prophesee’s sensor that Akida would be integrated into then Prophesee would be the one taking out an Akida licence; if Akida tech and the Prophesee sensor remain separate then Sony would probably be taking the licence and as it looks like they would probably go thru Megachips in that case. I can’t see Qualcomm being a part of this - they would only care about the resultant SoC in either case.

Unless they are obtaining the chips from Renesas who are due to have their chips rolling off the production line..
That way they could maintain privacy for competitive reasons.
If they obtained a licence from BRN it would have to be the subject of an ASX an announcement.
Renesas would be producing its AKIDA chips for its clients.
What client(s) is the question.

 

[equanimous]

Claim “uncertainty” about Akida 2.0 but don’t elaborate, just reference a previous MF article that also claims “uncertainty” but again doesn’t elaborate.

Lol these MF articles really are dumb as dogshit.

Solid Track record

The Motley Fool reviews

The Motley Fool (Forex or Share Trading Platform): 1.8 out of 5 stars from 517 genuine reviews on Australia's largest opinion site ProductReview.com.au.

www.productreview.com.au

 

[McHale]

Synsense

SynSense and Prophesee develop one-chip event-based smart sensing solution

Partnership leverages respective leadership in neuromorphic computing and sensing to realize on-sensor processing integration leading to small form-factor, low-power and cost-effective IoT solutions

www.prophesee.ai

Thanks @miaeffect, I knew it was syn..... something

 

[rgupta]

Amir Sherman on LinkedIn: Alif Semiconductor Demo of Battery-powered Embedded Vision with the…

One video explains all the entire story. By enabling a dedicated NPU accelerator for machine learning operations, Arm Ethos-U55, The Alif Semiconductor…

www.linkedin.com

 

[Diogenese]

https://www.diva-portal.org/smash/get/diva2:1779206/FULLTEXT01.pdf

Not sure if shared before. Decent study, but noted that its still early days. Suggested using other architectures that could perhaps yield more promising results than akida.

Hi JB49,

It's good to see Akida being used in University projects.

The comment that

"Investigation of other neural network architectures: Instead of using the AkidaNet ImageNet model, exploring different architectures could perhaps yield more promising results. Such architectures might include ResNet, Inception, or EfficientNet."

is not directed at the Akida SoC processor circuit. It is talking about the NN model.

The research question was:

"1.1 Research Question
How does the number of new images used, the number of neurons per class in the classification layer and the neuron weight affect the performance of edge learning of a spiking neural network implemented on BrainChip’s AKD1000 neuromorphic processor?"

The project was examining the effect of variations on the configuration and model parameters on the performance of Akida 1000.

MobilNet is an open source CNN model. AkidaNet is Brainchip's adaptation of Mobilnet to run on Akida's 4 bit NPUs.

PvdM's "4 Bits are Enough" explains AkidaNet.

https://brainchip.com/4-bits-are-enough/ :

“Table 1 provides the accuracy of several 4-bit CNN networks comparable to floating-point accuracies. For example, AkidaNet is a version of Mobilnet optimized for 4-bit classification, and many other example networks can be downloaded from the Brainchip website. In the quantization column below, ‘a’/’b’/’c’ where ‘a’ means weights bits for first layer, ‘b’ means weights bits for subsequent layers, ‘c’ means output activation map bits for every layer.

 

[wilzy123]

Well Brainchip tapes out akd1500 chip means that the chip has been through a lot of theoretical tests and it has passed. ie there isn't much risk about actually making them. But there is that last step. Brainchip could spend $4mill and produce a pile of them and hand them out. So MF are casting shade on Brainchip for not wasting $4mill in this exercise which isn't necessary if Brainchip know there are already other OEM's going to manufacture the chip.

Which articles did you write?

 

[Diogenese]

Thanks @miaeffect, I knew it was syn..... something

Was it the original syn?

 

[McHale]

Was it the original syn?

Well they were synning before BRN got involved in the story.

But if the prophesee is correct it will BRN who is winning.

 

[TopCat]

I haven’t seen this before and I’m not sure of the date this might have been released, but this is from Infineon’s Development Center in Dresden.

Development Center - Infineon Technologies

www.infineon.com

 

[Tothemoon24]

Great read

Unleashing Precision with Neuromorphic Vision in Robotic Drilling Systems - Khalifa University

By leveraging bio-inspired vision sensors and intelligent control algorithms, a new system can perceive the environment with exceptional clarity and make real-time adjustments to ensure accurate robotic drilling operations. Robotic drilling systems play a crucial role in various industries...

www.ku.ac.ae

The real-time responsiveness of the neuromorphic vision sensors enables the robotic drilling systems to adapt swiftly to changes in the drilling environment. This dynamic adaptability is particularly valuable in scenarios where the drilling targets were not static or when unexpected obstacles emerged during the drilling process. By continuously analyzing the visual data and making rapid adjustments, the robotic drilling system is able to maintain its precision and ensure optimal drilling outcomes.

 

[Diogenese]

Just saw a video of this on my news feed, but can't find it now.

Aria Research

Aria is an Australian company, looking for "sophisticated" investors, so not on ASX yet.

It is for echolocation glasses for vision impaired people.

It uses video object recognition with framing boxes and tells the user what the object is and also identifies its location.

They boast an all-day battery.

It collaborates with Sydney Uni, UTS, DISER, and a couple of others.

They are looking to recruit for clinical trials.

 

[AusEire]

Just saw a video of this on my news feed, but can't find it now.

Aria Research

Aria is an Australian company, looking for "sophisticated" investors, so not on ASX yet.

It is for echolocation glasses for vision impaired people.

It uses video object recognition with framing boxes and tells the user what the object is and also identifies its location.

They boast an all-day battery.

It collaborates with Sydney Uni, UTS, DISER, and a couple of others.

They are looking to recruit for clinical trials.

Saw a story on these guys on 9 news earlier. Maybe that's where you saw it?

 

[Diogenese]

Saw a story on these guys on 9 news earlier. Maybe that's where you saw it?

Yes. Thanks AE.

I found a US patent for a company called ARIA Tech:


US10078971B2 Media streaming methods, apparatus and systems 20140903

It seems very close, if antique technology from 2014.

Given their claimed 1 day battery life, I did mention Akida to Aria.

Techniques and systems are provided for assisting a user through a guidance mode activity. Such a user may be visually or otherwise impaired, or a user requiring help for other reasons. In some embodiments, a client device, held or worn by the user, may comprise a video capture device and several sensors, and may send video data and sensor data to a server. The server may comprise a processor and artificial intelligence. The server may send the video data and sensor data to an agent device. The agent device may provide content for display on an agent interface. An agent may view the agent interface, and assist the user in real time through audio instructions or other feedback.

[0121] For example, algorithms based on convolutional neural networks can be applied to video data feed received from the client device. A determination can be made as to whether the video data feed contains a specific feature, object, or activity. Such features, objects, or activities can be specified to the algorithm as predefined input. In other aspects, features, objects, and activities may be learned.

The principles of Aria Research are US entrepreneurs.

 

[jtardif999]

No relevance to our AKIDA.

The word AKIDA, in Brainchip AKIDA comes from the Greek word meaning spike (ie as in SNN - "Spiking" Neural Network).

In conrtext, to African and/or islamic cultures, there is also a word called Akida, however it has a meaning of certain, firm, etc.

Therefore, there is no link between the 2.


For those interested, Akida is also an African name, such as in Esse Akida. She is a Kenyan soccer player who plays in onr of the major leagues in Europe.

Maybe BRN should sponsor her

 

[jtardif999]

Well they were synning before BRN got involved in the story.

But if the prophesee is correct it will BRN who is winning.

So syners are not necessarily winners