BRN Discussion Ongoing

[Diogenese]

Exciting technology @TopCat

No I had not seen this article or picked up on this research with Xilinx.

There are clear links to Xilinx as you have shown and others have explored dots across a number of areas but no one has yet proven exactly what might be happening between Brainchip and Xilinx.

@Diogenese might be able to find a patent that could enhance the collective knowledge on this robotic arm being controlled by neural impulse.

My opinion only DYOR
FF

AKIDA BALLISTA

Hi @TopCat, FF,

I haven't read the article, but, if it is about using spikes to control an arm, that is a curious case of morphic resonance, as I had just referred to the sensor-to-brain path, which Akida does, and the brain-to-muscle path, which Akida does not do (yet).

As to Xilinks, we used their FPGA to make the BrainChip Accelerator which made a brief appearance here a couple of days ago.

 

[Violin1]

Ill pose the questions for everyone here now as It would be interesting to hear everyones response.

Im curious from your point of view, what are you most excited about with Akida and in what applications?

Is there a use case that hasnt been seen to date which you think could really benefit from Akida?

Do you have any thoughts on future editions of Akida and its potential?

Hi @equanimous
1. Most excited about the power saving aspects, particularly if it gets used in data centres and reduces data centre growth. This has the potential to contribute to power savings that benefit everyone on the planet, not just those who can afford a gizmo with Akida in it.
2. We haven't been privy to specific agricultural case uses so far. Plenty of speculation and connections but there has to be a heap of uses from rapid sensing of cargo/baggage at the border (to detect pests), on-farm precision techniques for planting, fertilising, pest management etc, then harvesting, sorting, packaging and through to transportation to alert on cool storage issues (such as in refrigerated containers). List goes on.
3. Future editions seem only limited by PVMD's available time! We must be looking at an expanding range of research and development that builds off success to date. The first fifteen years would have been sort of linear in the development but since Akida1000 they have a real base to work from. We already had the example of development being delayed to incorporate feedback from EAP/s but once it starts getting more broadly picked up then presumably there will be ideas being fed back to the company for enhancements/developments.

Oooh. You've gone and got me excited now - and I'd only just calmed down from The Pies' Big Win Last Night!!!!

 

[Bravo]

I was just reading this recent article about autonomous drones and law enforcement and I nearly chocked on my cuppa when I read this highlighted sentence. Move over Alexa and Siri, there's a new voice assistant in town!

Can AI drones help protect officers in these dangerous times?

Autonomous devices have the potential to be a gamechanger for law enforcement safety and reconnaissance

www.police1.com

 

[Fact Finder]

I was just reading this recent article about autonomous drones and law enforcement and I nearly chocked on my cuppa when I read this highlighted sentence. Move over Alexa and Siri, there's a new voice assistant in town!




View attachment 16367

Can AI drones help protect officers in these dangerous times?

Autonomous devices have the potential to be a gamechanger for law enforcement safety and reconnaissance

www.police1.com

Hi @Bravo

So would it be fair to say that you see some real significance in Brainchip releasing that video recently demonstrating an AKIDA voice controlled DRONE or am I reading toooo much into your post.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Fact Finder]

Hi @Bravo

So would it be fair to say that you see some real significance in Brainchip releasing that video recently demonstrating an AKIDA voice controlled DRONE or am I reading toooo much into your post.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[H.E. Pennypacker]

 

[Deena]

Deena,

The internet as we know it was rolled out in Australia in 1995. Corporate and Govt where the 1st big adopters who had access to fibre loops along with some small retail/household early adopters who put up with dial-up speeds. It really didn't take off in the broader community until around 2002 after Y2K thanks to people like yourself when it gained traction in the schools when kids came home pestering their parents to get it.

Before 1995, Corporate and Govt had their own routers and servers for internal use for their own WAN/LAN networks.

Thanks Proga. Edited my post. At my age the memory is not so good.

 

[Fact Finder]

This article from Accenture in places seems very familiar and indeed some of the language I feel confident has made its way into some recent material out of Brainchip. If someone has provided the link before I apologise but just reading it now after @Bravo’s post
listing Mercedes Benz voice with Alexa and Siri caused the following to jump out despite the absence of any reference to Brainchip:

“CASE STUDIES
Applied today:
Responsive voice control for smart vehicles
Owners of smart vehicles have high expectations: they want functions like self-parking and summoning features, but they also want natural, seamless interfaces to control their interactions with the car. Voice-based controls, along with implicit intent recognition, can provide greater personalization and smoother interactions. They also help unclutter the car’s dashboards and displays.
Why don’t today’s vehicles make wider use of this approach? Because it requires intensive computation. Conventional AI hardware is too power- hungry to run onboard continuously without draining the battery, and using only cloud-based AI and wake words creates too much lag, leading to a poor experience.
Neuromorphic technologies make efficient onboard AI possible. In a recent collaboration with an automotive client, we demonstrated that spiking neural networks running on a neuromorphic processor can recognize simple voice commands up to 0.2 seconds faster than a commonly used embedded GPU accelerator, while using up to a thousand times less power. This brings truly intelligent, low latency interactions into play, at the edge, even within the power-limited constraints of a parked vehicle”

https://www.accenture.com/_acnmedia/PDF-145/Accenture-Neuromorphic-Computing-POV.pdf

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Diogenese]

Hi @Diogenese

The following paper proposing the system being used here is where all the secrets seem to lie so if you feel inclined could you run your eye over this and provide your thoughts as to whether they need an IP licence from Brainchip to put their thoughts into action:

https://www.researchgate.net/publication/352150855_Toward_Robust_Cognitive_3D_Brain-Inspired_Cross-Paradigm_System/fulltext/60d5ef8e92851ca94487d038/Toward-Robust-Cognitive-3D-Brain-Inspired-Cross-Paradigm-System.pdf?origin=publication_detail

Regards
FF

AKIDA BALLISTA

Hi FF,

This paper is mainly directed to using 3D IC technology (invented by Samsung?) to produce a facsimile of the 3D layout of the brain, shortening the electrical paths between neurons.

Instead of using a 2D array of Multiple-AccumulationUnits (MACs), the authors converted to the 3D structure by using either through-silicon-vias (TSVs) or monolithic intertier vias (MIVs). Previously, we have explored the ability to integrate SNN into 3D-NoCs (Dang and Ben Abdallah, 2019; Ikechukwu et al., 2021). Instead of using 2D-NoCs, we extend the NoC to the third dimension, allowing the conventional neuron cluster design to work with small changes in the routing mechanism. As most of the state-of-the-art works still focus on dealing with 2DICs neuromorphic systems, we observe that exploring to 3D-ICs approach can bring several benefits such as smaller footprints and small number of hops.

They use 16-bit data and MACs and leaky-integrate-and-fire, so not N-of-M coding.

Instead of using a 2D array of Multiple-Accumulation Units (MACs), the authors converted to the 3D structure by using either through-silicon-vias (TSVs) or monolithic intertier vias (MIVs). [#### A "via" is the hole or tunnel providing a conductive path in a PCB or silicon chip ###]

As to determining infringement of any BRN patents, this is a "techno-semantic" exercise which involves item-by-item comparison between the infringement and the individual items in the claims of the patent, and their working interrelationship.

 

[Fredsnugget]

Hi @equanimous
1. Most excited about the power saving aspects, particularly if it gets used in data centres and reduces data centre growth. This has the potential to contribute to power savings that benefit everyone on the planet, not just those who can afford a gizmo with Akida in it.
2. We haven't been privy to specific agricultural case uses so far. Plenty of speculation and connections but there has to be a heap of uses from rapid sensing of cargo/baggage at the border (to detect pests), on-farm precision techniques for planting, fertilising, pest management etc, then harvesting, sorting, packaging and through to transportation to alert on cool storage issues (such as in refrigerated containers). List goes on.
3. Future editions seem only limited by PVMD's available time! We must be looking at an expanding range of research and development that builds off success to date. The first fifteen years would have been sort of linear in the development but since Akida1000 they have a real base to work from. We already had the example of development being delayed to incorporate feedback from EAP/s but once it starts getting more broadly picked up then presumably there will be ideas being fed back to the company for enhancements/developments.

Oooh. You've gone and got me excited now - and I'd only just calmed down from The Pies' Big Win Last Night!!!!

I am hoping Akida will be put into good use in medical diagnosis. As great as the multitude of applications are i think it finding patterns in medical imaging and other tests so much faster than the human eye can will be life changing for many. Having my grandson in ICU at present in an induced coma, and all the doctors doing so many tests, and still can only come up with some sort of brain infection. If only Akida could scan all the CT and MRI scans accurately and quickly a diagnosis might be made quicker than a dozen doctors arguing over there interpretation. One day hopefully

 

[Dhm]

I am hoping Akida will be put into good use in medical diagnosis. As great as the multitude of applications are i think it finding patterns in medical imaging and other tests so much faster than the human eye can will be life changing for many. Having my grandson in ICU at present in an induced coma, and all the doctors doing so many tests, and still can only come up with some sort of brain infection. If only Akida could scan all the CT and MRI scans accurately and quickly a diagnosis might be made quicker than a dozen doctors arguing over there interpretation. One day hopefully

I'm sure you have the vast family of the 1000 Eyes' sympathy and support, with love and prayers for your grandson and family.

 

[equanimous]

https://assets.researchsquare.com/files/rs-1860157/v1_covered.pdf?c=1658509150

 

[equanimous]

I am hoping Akida will be put into good use in medical diagnosis. As great as the multitude of applications are i think it finding patterns in medical imaging and other tests so much faster than the human eye can will be life changing for many. Having my grandson in ICU at present in an induced coma, and all the doctors doing so many tests, and still can only come up with some sort of brain infection. If only Akida could scan all the CT and MRI scans accurately and quickly a diagnosis might be made quicker than a dozen doctors arguing over there interpretation. One day hopefully

Sorry to hear that mate. Hopefully they find the root cause and required treatment early. In regards to your suggestion see below

Pathological identification of brain tumors based on the characteristics of molecular fragments generated by laser ablation combined with a spiking neural network​

SNN performed better than conventional machine learning methods for the analysis of similar and limited MFS information. With the ratio data type, the identification accuracy achieved 88.62% in 2 seconds​

Pathological identification of brain tumors based on the characteristics of molecular fragments generated by laser ablation combined with a spiking neural network

Quick and accurate diagnosis helps shorten intraoperative waiting time and make a correct plan for the brain tumor resection. The common cryostat section method costs more than 10 minutes and the diagnostic accuracy depends on the sliced and frozen process ...

www.ncbi.nlm.nih.gov

 

[Diogenese]

View attachment 16375

https://assets.researchsquare.com/files/rs-1860157/v1_covered.pdf?c=1658509150

Akida uses rank coding, not rate coding.

 

[equanimous]

Akida uses rank coding, not rate coding.

I wander why Chris Stevens and Todd Vierra like that post?

 

[JK200SX]

We get a mention in this article published today:

9 ASX-listed penny stocks on the move

A look at recent updates from 9 ASX-listed penny stocks - Kalkine Media

kalkinemedia.com

 

[Aretemis]

Hi @Diogenese

Me, myself and I along with my Blind Freddie believe that the AKIDA with LSTM IP has been in the hands of a select group of customers for some many months.

To come to this belief we have relied upon our deep scientific and technical knowledge amounting to zero and the following forensic known facts:

1. In June, 2019 Brainchip informed shareholder’s that AKIDA IP had been provided to a select group of early access customers - a large number of speculative dots now point to Valeo and Mercedes Benz being two of those customers.

2. It was not until 12 months later that Socionext received back the AKD1000 engineering samples from TSMC.

3. In November, 2021 Peter van der Made mentioned that the next version of AKD1000 with LSTM was shortly to be released to Engineering - Anil Mankar

4. At the AGM in May, 2022 the CEO Sean Hehir stated that he expected to be announcing the release of the IP in the second half 2022 and that if they decide to go to fab it would be first half 2023.

5. Peter van der Made has not been confined to Perth but has in recent weeks been helping out in the Paris offices with software and the US Offices with engineering both activities would seem unlikely if the AKIDA LSTM IP was still languishing incomplete on the bench in Perth. I note Paris is in Europe and Mercedes Benz and Valeo are there as well.

Now you may say I’m a dreamer but I am not the only one but I hope one day you will join us (Blind Freddie & me) and this idea will be as one.

My opinion only DYOR
FF

AKIDA BALLISTA

Well hopefully mrs karma will see us through

 

[Fact Finder]

I wander why Chris Stevens and Todd Vierra like that post?

Maybe the subtext of their 'likes' is that's right mate use the left door. Yep that's the one. The one on the left with the word Exit. Don't worry if its dark the lights automatic and once the door closes behind you it will come on. If it doesn't just keep going the way your going because there is no coming back from there.

 

[Izzzzzzzzzzy]

can’t wait for Christmas and brn being at $2

 

[Fact Finder]

We know that Mercedes Benz quantified the AKIDA efficiency as being 5 to 10 times that of any competitor product for their yet to be released Hey Mercedes offering.

This being so I believe it is logically available to say that this list of competitors that could not match up to AKIDA technology would have to include the known Mercedes Benz collaborators in the automotive space. It is just not logical to believe that Mercedes Benz would have ignored existing relationships in seeking the best available solution.

The following companies have been publicly acknowledged by Mercedes Benz as collaborating with them in the automotive space including the EQXX EV project:

Nvidia,
Google,
Samsung,
Intel,
LG,
Bosch,
Apple,
Valeo, and
Cadence.

By a strange coincidence all these listed companies claim to have some type of Ai advantage but given Mercedes Benz's statement it must not be in the area of Ai technology suitable for the Hey Mercedes voice control or in the Ai space that will allow a vehicle to think like you.

My opinion only DYOR
FF

AKIDA BALLISTA