BRN Discussion Ongoing

[thelittleshort]

“…Suryo developed a proof of concept drone-mounted animal tracker that can survey a large area and count the number of each type of animal that is present in it.

Rather than instrumenting each animal, this tracker uses computer vision and machine learning to locate and identify them. A single tracker can keep watch over an entire pasture full of animals.

And with the help of Edge Impulse, the algorithm runs entirely on an edge computing device onboard the drone — no cloud connection or subscription service is required.”

Link

 

[DingoBorat]

Musk approaches brain chip startup Synchron about deal amid Neuralink delays

Elon Musk has approached brain chip implant developer Synchron Inc for investment. His own company Neuralink plays catch-up in the race to connect the human brain.A Synchron spokesperson declined to comment.Elon Musk...

www.geo.tv

Did your heart skip a beat

It does show though, that Musk is man enough to admit his own venture Neuralink, isn't good enough and is willing to buy into something, that's getting there faster..

It could also apply, to his approach to ADAS.

 

[Diogenese]

View attachment 14539

“…Suryo developed a proof of concept drone-mounted animal tracker that can survey a large area and count the number of each type of animal that is present in it.

Rather than instrumenting each animal, this tracker uses computer vision and machine learning to locate and identify them. A single tracker can keep watch over an entire pasture full of animals.

And with the help of Edge Impulse, the algorithm runs entirely on an edge computing device onboard the drone — no cloud connection or subscription service is required.”

Link

Hi tls,

Back on the farm, we used to count the legs and divide by four.

This is a proof-of-concept arrangement, so not optimized for commercial exploitation. It is at this stage an algorithm running in Raspberry Pi.

The heat sink indicates the processor is generating a bit of heat. I'm not sure if they could dispense with the heat sink if they had Akida, but they could use a smaller one, which would lighten the load for the drone, and, as we know, Akida would extend the battery charge.

Apparently Edge impulse does a lot of this proof-of-concept modelling:

https://www.edgeimpulse.com/blog/getting-more-cycles-per-second-with-fomo

Note that this system does not appear to have one-shot learning, and needs a significant training database.

Basically they are using Edge Impulse's FOMO (Faster Objects, More Objects) algorithm:

https://www.edgeimpulse.com/blog/announcing-fomo-faster-objects-more-objects

It appears that there are synergies in using FOMO and Akida.

I hadn't previously looked into EI's FOMO, but the above link provides an excellent introduction, and the synergies become apparent.

FOMO uses a form of CNN and has significant limitations.

 

[equanimous]

Neuromorphic Engineering Integration from Materials and Devices to System and Algorithm

Conventional computing systems have been facing challenges in terms of energy efficiency and massive data processing when handling big data in various applications. The challenges mainly come from the von Neumann architecture-based processor used in the systems where memory and the central...

www.frontiersin.org

 

[equanimous]

Combinatorial optimization solving by coherent Ising machines based on spiking neural networks​

Bo Lu, Yong-Pan Gao, Kai Wen, Chuan Wang

Spiking neural network is a kind of neuromorphic computing which is believed to improve on the level of intelligence and provide advabtages for quantum computing. In this work, we address this issue by designing an optical spiking neural network and prove that it can be used to accelerate the speed of computation, especially on the combinatorial optimization problems. Here the spiking neural network is constructed by the antisymmetrically coupled degenerate optical parametric oscillator pulses and dissipative pulses. A nonlinear transfer function is chosen to mitigate amplitude inhomogeneities and destabilize the resulting local minima according to the dynamical behavior of spiking neurons. It is numerically proved that the spiking neural network-coherent Ising machines has excellent performance on combinatorial optimization problems, for which is expected to offer a new applications for neural computing and optical computing.

Comments:	5 pages, 4 figures, comments are welcome
Subjects:	Quantum Physics (quant-ph); Neural and Evolutionary Computing (cs.NE)
Cite as:	arXiv:2208.07502 [quant-ph]
	(or arXiv:2208.07502v1 [quant-ph] for this version)
	https://doi.org/10.48550/arXiv.2208.07502 Focus to learn more

Submission history​

From: Chuan Wang [view email]
[v1] Tue, 16 Aug 2022 02:19:05 UTC (1,024 KB)

Combinatorial optimization solving by coherent Ising machines based on spiking neural networks

Spiking neural network is a kind of neuromorphic computing that is believed to improve the level of intelligence and provide advantages for quantum computing. In this work, we address this issue by designing an optical spiking neural network and find that it can be used to accelerate the speed...

arxiv.org

 

[alwaysgreen]

Musk approaches brain chip startup Synchron about deal amid Neuralink delays

Elon Musk has approached brain chip implant developer Synchron Inc for investment. His own company Neuralink plays catch-up in the race to connect the human brain.A Synchron spokesperson declined to comment.Elon Musk...

www.geo.tv

Did your heart skip a beat

It does show though, that Musk is man enough to admit his own venture Neuralink, isn't good enough and is willing to buy into something, that's getting there faster..

It could also apply, to his approach to ADAS.

Geez wouldn't it be nice. A little offer of 6 billion to rocket our share price to $3 or $4. Our founders urge us to not accept because with everything in the pipeline, we are worth so much more, driving our share price to $6.

A man can dream...

 

[Diogenese]

Combinatorial optimization solving by coherent Ising machines based on spiking neural networks​

Bo Lu, Yong-Pan Gao, Kai Wen, Chuan Wang

Comments:	5 pages, 4 figures, comments are welcome
Subjects:	Quantum Physics (quant-ph); Neural and Evolutionary Computing (cs.NE)
Cite as:	arXiv:2208.07502 [quant-ph]
	(or arXiv:2208.07502v1 [quant-ph] for this version)
	https://doi.org/10.48550/arXiv.2208.07502 Focus to learn more

Submission history​

From: Chuan Wang [view email]
[v1] Tue, 16 Aug 2022 02:19:05 UTC (1,024 KB)

Combinatorial optimization solving by coherent Ising machines based on spiking neural networks

Spiking neural network is a kind of neuromorphic computing that is believed to improve the level of intelligence and provide advantages for quantum computing. In this work, we address this issue by designing an optical spiking neural network and find that it can be used to accelerate the speed...

arxiv.org

"Ising machines": so they're using SNNs in the cake shop now!

 

[thelittleshort]

Hi tls,

Back on the farm, we used to count the legs and divide by four.

This is a proof-of-concept arrangement, so not optimized for commercial exploitation. It is at this stage an algorithm running in Raspberry Pi.

The heat sink indicates the processor is generating a bit of heat. I'm not sure if they could dispense with the heat sink if they had Akida, but they could use a smaller one, which would lighten the load for the drone, and, as we know, Akida would extend the battery charge.

Apparently Edge impulse does a lot of this proof-of-concept modelling:

https://www.edgeimpulse.com/blog/getting-more-cycles-per-second-with-fomo

Note that this system does not appear to have one-shot learning, and needs a significant training database.

Basically they are using Edge Impulse's FOMO (Faster Objects, More Objects) algorithm:

https://www.edgeimpulse.com/blog/announcing-fomo-faster-objects-more-objects

It appears that there are synergies in using FOMO and Akida.

I hadn't previously looked into EI's FOMO, but the above link provides an excellent introduction, and the synergies become apparent.

FOMO uses a form of CNN and has significant limitations.

Thanks for feedback @Diogenese
Could it relate to the recent initiative? Or am I off course?

Edge Impulse Releases Deployment Support for BrainChip Akida Neuromorphic IP​

Edge Impulse, the leading platform for enabling ML at the edge, and BrainChip, the leading provider of neuromorphic AI IP technology, announced support for deploying Edge Impulse projects using the BrainChip Akida development kit.

Edge Impulse enables developers to rapidly build enterprise-grade ML algorithms, trained on real sensor data, in a low to no code environment.

These trained algorithms can now be quantized, optimized and converted to Spiking Neural Networks (SNN), which are compatible and can be deployed with BrainChip Akida™ devices.

This capability is available for new and existing Edge Impulse projects by using the BrainChip MetaTF model deployment block integrated on the platform. This deployment block enables free-tier developers and enterprise developer users to create and validate neuromorphic models for real-world use-cases and deploy on BrainChip Akida development kits.

 

[Diogenese]

Thanks for feedback @Diogenese
Could it relate to the recent initiative? Or am I off course?

Edge Impulse Releases Deployment Support for BrainChip Akida Neuromorphic IP​

Edge Impulse, the leading platform for enabling ML at the edge, and BrainChip, the leading provider of neuromorphic AI IP technology, announced support for deploying Edge Impulse projects using the BrainChip Akida development kit.

Edge Impulse enables developers to rapidly build enterprise-grade ML algorithms, trained on real sensor data, in a low to no code environment.

These trained algorithms can now be quantized, optimized and converted to Spiking Neural Networks (SNN), which are compatible and can be deployed with BrainChip Akida™ devices.

This capability is available for new and existing Edge Impulse projects by using the BrainChip MetaTF model deployment block integrated on the platform. This deployment block enables free-tier developers and enterprise developer users to create and validate neuromorphic models for real-world use-cases and deploy on BrainChip Akida development kits.

Yes. When EI's FOMO is supercharged with Akida, the G-forces will be overpowering.

The synergies of combining EI's FOMO, which uses CNN, with Akida, which can convert CNN to SNN, enabling N-of-M coding greatly reduces the compute workload, and also brings one-shot learning into play.

This shows the wisdom of incorporating CNN2SNN into Akida. CNN is the pre-existing standard with which many designers working in the field will be comfortable, but now CNN applications can be converted to SNN.

EI's FOMO can be adapted to locate objects, rather than classifying them, and thus uses centroids instead of bounding boxes derived from the input signals.

https://www.edgeimpulse.com/blog/announcing-fomo-faster-objects-more-objects

"Design choices
Trained on centroid

The main design decisions for FOMO are based on the idea that a lot of object detection problems don't actually need the size of the objects but rather just their location in the frame. Once we know the location of salient objects we can ask further questions such as "Is an object above/below another?" or "How many of these objects are in view?"

Based on this observation, classical bounding boxes are no longer needed. Instead, a detection based on the centroids of objects is enough.

Note that, to keep the interoperability with other models, your training image input still uses bounding boxes."

Even so, they only manage 10 frames per second with their smallest version of FOMO, which they considered "amazingly fast".

"Designed to be small and amazingly fast

One of the first goals, when we started to design FOMO, was to run object detection on microcontrollers where flash and RAM are most of the time very limited. The smallest version of FOMO (96x96 grayscale input, MobileNetV2 0.05 alpha) runs in <100KB RAM and ~10 fps on a Cortex-M4F at 80MHz."

The combination of Akida with EI's FOMO will blow their socks off, even doing both classification and location. This greatly expands the potential field of use of EI's FOMO.

 

[Meetupsoon]

BrainChip selected by U.S. Air Force Research Laboratory to develop AI-based radar​

2022-08-21 04:13 HKT

https://inf.news/en/science/791bf988657af0e134c475ca748067c6.html

 

[Cyw]

BrainChip selected by U.S. Air Force Research Laboratory to develop AI-based radar​

2022-08-21 04:13 HKT

https://inf.news/en/science/791bf988657af0e134c475ca748067c6.html

Wonder what kind of money we are talking about in this type of deal. There may be a licence sold but no royalties.

 

[Fox151]

Wonder what kind of money we are talking about in this type of deal. There may be a licence sold but no royalties.

Wasn't this something we already knew?

 

[Cyw]

Wasn't this something we already knew?

I might have missed it. Can someone post a link or reference?

 

[Townyj]

This has been put out a few times already.

Information Systems Labs Joins BrainChip Early Access Program

BrainChip announced that Information Systems Labs is developing an AI-based radar research solution for the AFRL based on Akida™

brainchip.com

Information Systems Labs Joins BrainChip Early Access Program

LAGUNA HILLS, Calif., January 09, 2022--Information Systems Laboratories, is developing an AI-based radar research solution for the Air Force Research Laboratory (AFRL) based on Akida

au.finance.yahoo.com

 

[TopCat]

Not saying this is Akida, but absolutely amazing technology to watch. Woolworths is to open one in Sydney by 2024. Think of the power savings they could get by using Akida!

 

[Potato]

Can i get an Akida Balista please?

 

[Stable Genius]

Liked by RT.

A bit technical for me but I understood it to mean things are progressing with Si-Five and there are big opportunities in the future!

Redirecting

Enjoy

 

[Diogenese]

Can i get an Akida Balista please?

I'm waiting for the coffee machine.

 

[Iseki]

Liked by RT.

A bit technical for me but I understood it to mean things are progressing with Si-Five and there are big opportunities in the future!

Redirecting

Enjoy

On the technical front here is my take
. Because we are trying to take over the edge, Akida's performance is not measured so much in TOPS, but TOPS/W (ie per watt.)
. We really want Akida IP to be tied into their smallest chip, ( so just an MCU), so that this can be placed adjacent to each and every sensor.
( so the SiFive E2 would be best choice.)
. This Akida informed MCU will pass on just the information required to whatever vector based trained neural network is appropriate.
. So What I'm praying for is an ann that says "SuchandSuch SoC developer publicly announces their adoption of SiFive and Akida IP to produce worlds smallest and most versatile sensor MCU"
. So the question is, which sensor manufacturer apart from Renesas who seems to have gone down the Arm road, is big enough to push for this?

 

[MADX]

For those who understand the tech, you will find the link below interesting. It is from the Weebit nano forum. Along with slymeat we see increasing cross-tech emerging between WBT and BRN. I believe BRN holders should keep an eye on WBT.

WBT Discussion 2022

"This work was supported in part by SERB-CRG/2018/001901 and CYRAN AI Solutions. ... and 2) A. Bricalli, G. Piccolboni, G. Molas and A. Regev are with Weebit Nano, Hod Hasharon, Israel (e-mail: amir@weebit-nano.com ). ∗S. K. Kingra and V. Parmar contributed equally for the manuscript." 4 Aug...

thestockexchange.com.au