BRN Discussion Ongoing

  • Like
Reactions: 4 users
As a reward for our good behaviour I have just brought over from the NASA thread the following extracts from an SBIR software contract which Rocket had linked and challenged me to highlight anything of interest. Well the following is my take on what was interesting about this SBIR.

It is in my opinion pretty amazing stuff when you consider they mention Loihi and True North in the full knowledge that they cannot meet the specified requirements under COTS and SWaP. (PS: Somewhere I read something about Brainchip and open sourced hardware recently.) The link to the full document is at the end.
My opinion only DYOR
FF

AKIDA BALLISTA

DESCRIPTION: Conventional computing architectures are running up against a quantum limit in terms of transistor size and efficiency, sometimes referred to as the end of Moore’s Law. To regain our competitive edge, we need to find a way around this limit. This is especially relevant for small size, weight, and power (SWaP)-constrained platforms. For these systems, scaling Von Neumann computing becomes prohibitively expensive in terms of power and/or SWaP.Biologically inspired neural networks provide the basis for modern signal processing and classification algorithms.

Implementation of these algorithms on conventional computing hardware requires significant compromises in efficiency and latency due to fundamental design differences. A new class of hardware is emerging that more closely resembles the biological neuron model, also known as a spiking neuron model; mathematically describing the systems found in nature and may solve some of these limitations and bottlenecks. Recent work has demonstrated performance gains using these new hardware architectures and have shown equivalence to converge on a solution with the same accuracy [Ref 1].The most promising of the new class are based on Spiking Neural Networks (SNN) and analog Processing in Memory (PiM) where information is spatially and temporally encoded onto the network. It can be shown that a simple spiking network can reproduce the complex behavior found in the neural cortex with significant reduction in complexity and power requirements [Ref 2]……

“It is recommended to use open source languages, software, and hardware when possible.”…

PHASE III:
Refine algorithms and test with hardware. Validate models with data provided by Naval Air Warfare Center (NAWC) Aircraft Division (AD)/Weapons Division (WD). Transition model to the warfare centers. Development of documentation, training manuals, and software maintenance may be required.Heavy commercial investments in machine learning and artificial intelligence will likely continue for the foreseeable future. Adoption of hardware that can deliver on orders of magnitude in SWaP performance for intelligent mobile machine applications is estimated to be worth 10^9-10^12 global dollars annually.) Provide the software tools needed to optimize the algorithms and hardware integration. This effort would be a significant contribution to this requirement. Industries that would benefit from successful technology development include automotive (self-driving vehicles), personal robots, and a variety of intelligent sensors.

KEYWORDS: Spiking Neural Network, Neuromorphic Computing, Modeling, Convolution Neural Network, Analog Memory, Processing in Memory

https://www.sbir.gov/node/1696107
 
  • Like
  • Fire
  • Love
Reactions: 28 users

Slade

Top 20
This presentation is from a SiFive field engineer at the Semisrael tech webinar two days ago. The presentation is about SiFive's RISC-V, which we know BrainChip will be a part of. This video is for the tech heads, or those whose favourite movie is Zulu. I myself wasn't able to get much from it. The presenter looks like he would be about as much fun at a party as my father in law.

 
  • Like
  • Haha
  • Fire
Reactions: 18 users
This presentation is from a SiFive field engineer at the Semisrael tech webinar two days ago. The presentation is about SiFive's RISC-V, which we know BrainChip will be a part of. This video is for the tech heads, or those whose favourite movie is Zulu. I myself wasn't able to get much from it. The presenter looks like he would be about as much fun at a party as my father in law.


Sorry just tried I need Michael Caine to read for his part. Leave it for someone else.

FF.
 
  • Haha
  • Like
  • Fire
Reactions: 6 users
Sorry just tried I need Michael Caine to read for his part. Leave it for someone else.

FF.
happy birthday film GIF
 
  • Haha
  • Like
Reactions: 3 users
As a reward for our good behaviour I have just brought over from the NASA thread the following extracts from an SBIR software contract which Rocket had linked and challenged me to highlight anything of interest. Well the following is my take on what was interesting about this SBIR.

It is in my opinion pretty amazing stuff when you consider they mention Loihi and True North in the full knowledge that they cannot meet the specified requirements under COTS and SWaP. (PS: Somewhere I read something about Brainchip and open sourced hardware recently.) The link to the full document is at the end.
My opinion only DYOR
FF

AKIDA BALLISTA

DESCRIPTION: Conventional computing architectures are running up against a quantum limit in terms of transistor size and efficiency, sometimes referred to as the end of Moore’s Law. To regain our competitive edge, we need to find a way around this limit. This is especially relevant for small size, weight, and power (SWaP)-constrained platforms. For these systems, scaling Von Neumann computing becomes prohibitively expensive in terms of power and/or SWaP.Biologically inspired neural networks provide the basis for modern signal processing and classification algorithms.

Implementation of these algorithms on conventional computing hardware requires significant compromises in efficiency and latency due to fundamental design differences. A new class of hardware is emerging that more closely resembles the biological neuron model, also known as a spiking neuron model; mathematically describing the systems found in nature and may solve some of these limitations and bottlenecks. Recent work has demonstrated performance gains using these new hardware architectures and have shown equivalence to converge on a solution with the same accuracy [Ref 1].The most promising of the new class are based on Spiking Neural Networks (SNN) and analog Processing in Memory (PiM) where information is spatially and temporally encoded onto the network. It can be shown that a simple spiking network can reproduce the complex behavior found in the neural cortex with significant reduction in complexity and power requirements [Ref 2]……

“It is recommended to use open source languages, software, and hardware when possible.”…

PHASE III:
Refine algorithms and test with hardware. Validate models with data provided by Naval Air Warfare Center (NAWC) Aircraft Division (AD)/Weapons Division (WD). Transition model to the warfare centers. Development of documentation, training manuals, and software maintenance may be required.Heavy commercial investments in machine learning and artificial intelligence will likely continue for the foreseeable future. Adoption of hardware that can deliver on orders of magnitude in SWaP performance for intelligent mobile machine applications is estimated to be worth 10^9-10^12 global dollars annually.) Provide the software tools needed to optimize the algorithms and hardware integration. This effort would be a significant contribution to this requirement. Industries that would benefit from successful technology development include automotive (self-driving vehicles), personal robots, and a variety of intelligent sensors.

KEYWORDS: Spiking Neural Network, Neuromorphic Computing, Modeling, Convolution Neural Network, Analog Memory, Processing in Memory

https://www.sbir.gov/node/1696107
I might have been a bit vague with why this is exciting to me.

Yesterday in one of the articles announcing the Brainchip SiFive partnering there was a three questions section and Kang answering on behalf of SiFive
stated that Brainchip was bringing hardened Ai IP to their platform.

NASA was paying to harden AKIDA.

SiFive’s platform is open sourced hardware and with Brainchip’s assistance it will offer a hardened SCNN integrated.

What other opened sourced hardware product of this type that is COTS and SWaP ready is there in the World particularly if you exclude from consideration all those countries that would not meet NASA/DARPA national security requirements.

Also I am excited about how much NASA thinks this will be worth commercially. I know I am shallow being excited about the huge financial rewards.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 39 users

Build-it

Regular
I believe Akida is still at the “far-edge”, to use a term coined by Rob Telson (I believe). Where the actual inputs are processed to differenciate the important stuff from the noise. And where micro- and milli-watts are consumed

The term “edge” has unfortunately seemed to shift a fair way away from the far-edge. It now encompasses not only the entire smart sensor but actuators to respond to sensor, mechanisms to perform an activity, and mechanisms to provide feedback up the chain - ultimately to a human operator . i.e. the entire system. Power consumption will still be important, but I dare say it might be slipping up into partial watts.

Maybe this edge is more the point from which the human operator no longer needs to care about what goes on under the bonnet. This definition of an edge could be applied to physical objects such as cameras, car dash boards and the like.

The SiFive—BrainChip partnership seems to be creating something that will satisfy this edge, and that’s Ok by me because that is an easy edge to market and for the world to relate to. The world loves a system that does something tangeable. I think the general consumer market struggles to recognise the advantages of a spiking neural network. So let’s stop confusing them.

A Polish colleague of mine, who learnt English as a second language and hence was VERY specific with grammar, would have said this edge is not “the edge” but rather is just “an edge”. There are many interfaces that can be called “an edge” and even “another edge” but there is only one that can be called “the edge”.

“The edge” is, in all reality, the “far-edge”! The point where the stimuli and initial responses exist!

To quote Brainchip,

We are pleased to partner with SiFive and have the opportunity to have our Akida technology integrated with their market-leading product offerings, creating an efficient combination for edge compute”

IMO an Edge device that can run on as little as 4 x AAA batteries that can travel to all corners of the globe and not require data is my best version. Particularly as a medical device.

Ofcourse this is not our only offering but one that is true beneficial AI at the edge.

As Stable posted a while back
Nanonose expected to report May 30


Edge Compute
 
  • Like
  • Fire
  • Love
Reactions: 27 users

Slade

Top 20
Sorry just tried I need Michael Caine to read for his part. Leave it for someone else.

FF.
I am watching Zulu in stages. Five sittings so far and I am up to the riveting scene below. Does it get any better than this!

1649315936741.png

I mean, does this get any better?
 
Last edited:
  • Haha
  • Like
  • Love
Reactions: 6 users
After am watching Zulu in stages. Five sittings so far and I am up to the riveting scene below. Does it get any better than this!

View attachment 3877
I mean, does this get any better?
I can’t say as a still shot of a face that features throughout does not set the scene but if you are watching it for entertainment you are in the wrong theatre. LOL

It is the underlying principles. It was actually shown as a training aid at Duntroon about fifty years ago. That is how I became aware of it initially.

FF.
 
  • Like
Reactions: 3 users

Slade

Top 20
I can’t say as a still shot of a face that features throughout does not set the scene but if you are watching it for entertainment you are in the wrong theatre. LOL

It is the underlying principles. It was actually shown as a training aid at Duntroon about fifty years ago. That is how I became aware of it initially.

FF.
I have to confess, I read the wiki about the movie first which puts it into perspective. I was interested to find out that it was filmed not so far from the historical location and that most of the extras are actual Zulu warriors. I wonder how they felt about acting in a movie where 1400 of them get beat by such a small foreign battalion.
 
  • Fire
  • Like
Reactions: 2 users
Isnt it a given that NASA be using AKIDA? They just haven't told us yet that it's a signed deal.. how could they not? 🤷‍♂️
Yes it is a given but what we know allowed for us to conclude possible uses in CubeSats and camera type sensors.

What we have here with SiFive at NASA could be a complete paradigm shift by NASA away from all its traditional partners and computer technology to RISC-5 Ai platforms.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
Reactions: 10 users
I have to confess, I read the wiki about the movie first which puts it into perspective. I was interested to find out that it was filmed not so far from the historical location and that most of the extras are actual Zulu warriors. I wonder how they felt about acting in a movie where 1400 of them get beat by such a small foreign battalion.
They probably did not tell them until the final scene. LOL FF
 
  • Haha
  • Like
Reactions: 3 users

Dhm

Regular
I might have been a bit vague with why this is exciting to me.

Yesterday in one of the articles announcing the Brainchip SiFive partnering there was a three questions section and Kang answering on behalf of SiFive
stated that Brainchip was bringing hardened Ai IP to their platform.

NASA was paying to harden AKIDA.

SiFive’s platform is open sourced hardware and with Brainchip’s assistance it will offer a hardened SCNN integrated.

What other opened sourced hardware product of this type that is COTS and SWaP ready is there in the World particularly if you exclude from consideration all those countries that would not meet NASA/DARPA national security requirements.

Also I am excited about how much NASA thinks this will be worth commercially. I know I am shallow being excited about the huge financial rewards.

My opinion only DYOR🤪
FF

AKIDA BALLISTA
I am publishing this because I didn't understand what 'hardened' AI means, and I thought I may not be the only one - though I could be - :unsure:

Screen Shot 2022-04-07 at 5.33.33 pm.png
 
  • Like
Reactions: 8 users
I am publishing this because I didn't understand what 'hardened' AI means, and I thought I may not be the only one - though I could be - :unsure:

View attachment 3878
We are talking two different things. He said “hardened” not “hard”:

What are hardened electronics?

Radiation-hardened electronics, also called rad-hard electronics, are electronic components (circuits, transistors, resistors, diodes, capacitors, etc.), single-board computer CPUs, and sensors that are designed and produced to be less susceptible to damage from exposure to radiation and extreme temperatures (-55°C to ...

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Love
Reactions: 16 users
D

Deleted member 118

Guest
Now does anybody else need more evidence to solidify there investment in Brainchip?

There is a volcano brewing and when this baby is about to erupt, look out!

 
  • Like
Reactions: 3 users

Dhm

Regular
We are talking two different things. He said “hardened” not “hard”:

What are hardened electronics?

Radiation-hardened electronics, also called rad-hard electronics, are electronic components (circuits, transistors, resistors, diodes, capacitors, etc.), single-board computer CPUs, and sensors that are designed and produced to be less susceptible to damage from exposure to radiation and extreme temperatures (-55°C to ...

My opinion only DYOR
FF

AKIDA BALLISTA
Thanks FF. Mind you, the definition I looked up did sound convincing.....
 
  • Love
Reactions: 1 users

Townyj

Ermahgerd
We are talking two different things. He said “hardened” not “hard”:

What are hardened electronics?

Radiation-hardened electronics, also called rad-hard electronics, are electronic components (circuits, transistors, resistors, diodes, capacitors, etc.), single-board computer CPUs, and sensors that are designed and produced to be less susceptible to damage from exposure to radiation and extreme temperatures (-55°C to ...

My opinion only DYOR
FF

AKIDA BALLISTA

Beat me to it FF
 
  • Haha
  • Like
Reactions: 2 users

Foxdog

Regular
Yes it is a given but what we know allowed for us to conclude possible uses in CubeSats and camera type sensors.

What we have here with SiFive at NASA could be a complete paradigm shift by NASA away from all its traditional partners and computer technology to RISC-5 Ai platforms.

My opinion only DYOR
FF

AKIDA BALLISTA
Unheard of in only the recent past but now with players like SpaceX in the market NASA needs to become more nimble to remain relevant - such a paradigm shift might be exactly what is going on.
 
  • Like
  • Love
Reactions: 5 users
We are talking two different things. He said “hardened” not “hard”:

What are hardened electronics?

Radiation-hardened electronics, also called rad-hard electronics, are electronic components (circuits, transistors, resistors, diodes, capacitors, etc.), single-board computer CPUs, and sensors that are designed and produced to be less susceptible to damage from exposure to radiation and extreme temperatures (-55°C to ...

My opinion only DYOR
FF

AKIDA BALLISTA
Just to be absolutely sure I am not relying upon a flawed memory I found the quote:

“VentureBeat: Does BrainChip use any of Arm’s IP in its chips? Arm is known for low-power and high performance.”

Kang: BrainChip has discussed Arm IP for their product line. Arm processors have built a reputation for low-power based on comparisons to x86-based products. SiFive Intelligence products compare well to Arm products through offering improved performance-per-area of up to 30% combined with a single ISA for simpler programming, and a modular approach that aligns well to working with hardened AI IP such as developed by BrainChip.”

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Love
Reactions: 14 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
Hardened as in radiation-hardened
Not hard as in complex
And definitely not hard as in 🍆

😚
 
Last edited:
  • Haha
  • Like
  • Love
Reactions: 18 users
Top Bottom