BRN Discussion Ongoing

[Rise from the ashes]

xTechSearch 7- Competitions - xTechSearch

www.arl.army.mil

 

[Rise from the ashes]

Advancements in technology are going at break neck speeds A.I is truly a wonderful thing.
Until it's not.

 

[Moonshot]

Speaking of science fiction… that’s how it starts.

 

[Foxdog]

25 of those years overlapped, if that makes sense, so let's reduce the 90 back down to 64 but thinking I'm still 30 when I see a very
good looking lady.

Yes, but keep selling the story, we have a fantastic company and the more Australians who get to know about us and our brilliant
technology the better, as far as I'm personally concerned, we will always be Australian, thanks to Peter choosing our country to
develop and create his dream.

60 is the new 30 tech, so you're travelling well

 

[Steve10]

FYI

Chart with US 2 year bond yield (blue line) & US interest rate (orange line).

The green vertical lines is when 20ema crosses under 100ema for US 2 year bond yields on daily chart. When it crosses under coincides with peak in US interest rate. It's due to cross under sometime next week. US Feds should pause at next FMOC meeting 31/1-1/2/2023.

The S&P500 chart is due to golden cross (50sma crosses above 200sma) around end of January. Nasdaq chart is due to golden cross sometime in 3rd week February & IWM by mid next week. DJI has already golden crossed.

The US government yesterday changed the way CPI is calculated. They were calculating CPI using 2019/2020 data with lower base & now they will be using 2021 data with higher base = lower YOY CPI.

By the end of Q1 CY23 the CPI should go down by around 3% & the US Feds may commence rate cuts.

JANUARY 2023 CPI WEIGHT UPDATE

• Starting with January 2023 data, the BLS plans to update weights annually for the Consumer Price Index based on a single calendar year of data, using consumer expenditure data from 2021. This reflects a change from prior practice of updating weights biennially using two years of expenditure data.

https://www.bls.gov/cpi/

 

[Moonshot]

https://www.eetimes.com/perceive-runs-transformers-at-the-edge-with-second-gen-chip/

I’m more worried about this sort of chip - traditional AI done faster - at the edge than Loihi in terms of competition.

 

[VictorG]

https://www.eetimes.com/perceive-runs-transformers-at-the-edge-with-second-gen-chip/

I’m more worried about this sort of chip - traditional AI done faster - at the edge than Loihi in terms of competition.

I'm probably the least tech savvy on this forum but I don't see Ergo2 capable of one shot or few shot learning, doesn't use spiking neuromorphic architecture and is 100 times more power hungry than Akida.
It's 7mm ×7mm so likely very expensive.

 

[supersonic001]

Perceive has added hardware support for transformations to its 100-mW edge AI chip, and even demonstrated RoBERTa inference back at CES. | Edge Impulse (a Qualcomm company)

Perceive has added hardware support for transformations to its 100-mW edge AI chip, and even demonstrated RoBERTa inference back at CES. cc: EE Times | Electronic Engineering Times, Sally Ward-Foxton, Steve Teig

www.linkedin.com

A SWF article from EE times fyi - do we or font we aspire to Transformers

 

[Moonshot]

I'm probably the least tech savvy on this forum but I don't see Ergo2 capable of one shot or few shot learning, doesn't use spiking neuromorphic architecture and is 100 times more power hungry than Akida.
It's 7mm ×7mm so likely very expensive.

Agree it doesn’t have on chip learning but they are claiming 30fps at 17mw whereas akida 1000 is 30 FPS @ 157 mw in 28 nm
Don’t know what nm Ergo is, maybe Dio can help clarify?
Edit: built on GlobalFoundries' 22FDX process node

For comparison, Renesas DRP can do 30 frames per second at 3.1 Watts

Perceive's Ergo 2 Offers Multi-Model On-Device Machine Learning in a Sub-100mW Power Envelope - Hackster.io

Offering a claimed four times the compute of the original, Perceive's latest system-on-chip comes with some impressive efficiency claims.

www.hackster.io

https://perceive.io/wp-content/uploads/2023/01/2022-02-11_Ergo-2-Product-Overview.pdf

Info on the process- same cost as 28nm

https://www.gsaglobal.org/wp-content/uploads/2012/05/22FDX_GSA_AMS_Webinar_Schaeffer_rev2-JSA-rev-7.pdf

By my reading the after accounting for process to produce the chip still makes them about twice as power efficient as akida. Need some expert help!

 

[stuart888]

Prophesee have a White Paper out on event based sensing, "Metavision for Machines". It probably isn't appropriate for me to publish it myself as I had to apply for it, but here is a small teaser for you
View attachment 27580

It sounds very exciting from where I am sitting. Lets hope Prophesee can get the word out to the world!

Yes, @Dhm! Prophesee is getting the word out, like upcoming Jan 25th TinyML Presentation!

Go Luca Verre, spread the word of the Brainchip Akida, sitting on the pole (love it @TECH)!

https://www.tinyml.org/event/tinyml-trailblazers-series-tinyml-success-stories-with-luca-verre

 

[Jasonk]

Akida inside?
Seems to tick a lot of boxes and involve know partners.

 

[VictorG]

Agree it doesn’t have on chip learning but they are claiming 30fps at 17mw whereas akida 1000 is 30 FPS @ 157 mw in 28 nm
Don’t know what nm Ergo is, maybe Dio can help clarify?
Edit: built on GlobalFoundries' 22FDX process node

For comparison, Renesas DRP can do 30 frames per second at 3.1 Watts

Perceive's Ergo 2 Offers Multi-Model On-Device Machine Learning in a Sub-100mW Power Envelope - Hackster.io

Offering a claimed four times the compute of the original, Perceive's latest system-on-chip comes with some impressive efficiency claims.

www.hackster.io

https://perceive.io/wp-content/uploads/2023/01/2022-02-11_Ergo-2-Product-Overview.pdf

Info on the process- same cost as 28nm

https://www.gsaglobal.org/wp-content/uploads/2012/05/22FDX_GSA_AMS_Webinar_Schaeffer_rev2-JSA-rev-7.pdf

By my reading the after accounting for process to produce the chip still makes them about twice as power efficient as akida. Need some expert help!

Agreed, this one is for @Diogenese

 

[Bravo]

Perceive has added hardware support for transformations to its 100-mW edge AI chip, and even demonstrated RoBERTa inference back at CES. | Edge Impulse (a Qualcomm company)

Perceive has added hardware support for transformations to its 100-mW edge AI chip, and even demonstrated RoBERTa inference back at CES. cc: EE Times | Electronic Engineering Times, Sally Ward-Foxton, Steve Teig

www.linkedin.com

A SWF article from EE times fyi - do we or font we aspire to Transformers

View attachment 27624

As I understand it, AKIDA 2000 is going to be optimized for LSTM and transformers. Any comparison to Perceive would be difficult at this stage as AKIDA 2000 has yet to be launched.

 

[stuart888]

10,000 fps equivalent. I take this to mean that Prophesee's DVS event camera can capture movement (change in light impinging on a pixel) detected by individual pixels at 10,000 Hz, not a full screen of pixels in a frame. DVS cameras do not have a shutter, so the photoreceptor plate is continuously exposed to the field of view. The thing which would limit the speed which a DVS could capture movement would be the response time of the pixels unencumbered by any frame rate - without the inherent delay of the fixed frame configuration of normal video. The pixels fire asynchronously as the light impinging on the individual pixels changes.

Going back to Prophesee's comments about Akida, Akida can accept asynchronous input from individual pixels. It does not need to wait for a full frame of image data. It is able to receive individual "events" as they occur. So it seems from Prophesee's comments, Akida is capable of matching the performance of the Prophesee DVS, something that frame-based system cannot do.

On the other hand, nViso has tested Akida with framed video to over 1.6 k fps.

The 10,000 fps equivalent did baffle my mind, plus the DVS cameras without a shutter!

Lots of answers to how this is done appears to be the fantastic SNN article by Brainchip 4 Bits are Enough, but I liked the Nature article on: neuromorphic computing algorithms and applications that yells Brainchip Akida. Cannot remember who posted it, but thanks, it really helps.

Highly parallel operation: neuromorphic operations are inherently parallel, where all of the neurons and synapses can potentially be operating simultaneously.

Simple Operations: the computations performed by neurons and synapses are relatively simple when compared with the parallelized von Neumann systems.

Collocated processing and memory: there is no notion of a separation of processing and memory in neuromorphic hardware. Although neurons are sometimes thought of as processing units and synapses are sometimes thought of as memory, the neurons and synapses both perform processing and store values in many implementations. The collocation of processing and memory helps mitigate the von Neumann bottleneck regarding the processor/memory separation, which causes a slowdown in the maximum throughput that can be achieved.

https://www.nature.com/articles/s43588-021-00184-y

 

[VictorG]

Agreed, this one is for @Diogenese

Not one for quoting ChatGPT but I use this instance as an exception.

Differentiate between Akida and Ergo 2.

Akida and Ergo 2 are both neuromorphic processors developed by different companies for a wide range of applications such as image and speech recognition, autonomous vehicles, and industrial automation. However, there are some key differences between the two processors.

Architecture: Akida is based on a spiking neural network architecture, which mimics the behavior of neurons in the brain. Ergo 2, on the other hand, is based on a more traditional artificial neural network architecture.

Power consumption: Akida is designed to be highly power-efficient, with low power consumption and high performance. Ergo 2 also claims to be low-power but the specific power consumption figures are not publicly available.

Programming: Akida provides a software development kit (SDK) and programming model that is designed to be easy to use, even for developers with limited experience in neuromorphic computing. Ergo 2 also has a development kit but the information about the programming model is not publicly available.

Pricing: Akida is available as a system-on-a-chip (SoC) and as a development board, with the SoC available for $1.65 in volume orders. Ergo 2 is available as a development kit that includes a development board and software tools, and is priced at $2,995.

Company: Akida is developed by Brainchip, a company that specializes in neuromorphic computing solutions. Ergo 2 is developed by Mythic, a company that specializes in low-power artificial intelligence processors.

In summary, while both Akida and Ergo 2 are neuromorphic processors designed for a wide range of applications, they differ in terms of architecture, power consumption, programming, pricing and the company that developed them.

 

[Serengeti]

Happy Saturday everyone!

Came across a number of AI podcasts and thought I’d share. Not directly linked to BRN, they discuss AI in general but thought I would post for those who may be interested.

Podcast: The future of artificial intelligence with tech CEO Kai-Fu Lee

Listen: Artificial intelligence is changing the way we live — and very soon it'll go beyond medical breakthroughs and the algorithms that control your social newsfeeds. Will AI become the biggest technological disrupter since the Industrial Revolution, replacing workers with robots? On this...

www.gzeromedia.com

AI at the tipping point: danger to information, promise for creativity

Artificial intelligence is on everyone's mind these days. The potential for AI to mess up democracy is scary, but the truth is that it can also make the world a better place. So, are bots good or bad for us? We asked a few experts to weigh in during the Global Stage livestream conversation...

www.gzeromedia.com

 

[misslou]

Not one for quoting ChatGPT but I use this instance as an exception.

Differentiate between Akida and Ergo 2.

Akida and Ergo 2 are both neuromorphic processors developed by different companies for a wide range of applications such as image and speech recognition, autonomous vehicles, and industrial automation. However, there are some key differences between the two processors.

Architecture: Akida is based on a spiking neural network architecture, which mimics the behavior of neurons in the brain. Ergo 2, on the other hand, is based on a more traditional artificial neural network architecture.

Power consumption: Akida is designed to be highly power-efficient, with low power consumption and high performance. Ergo 2 also claims to be low-power but the specific power consumption figures are not publicly available.

Programming: Akida provides a software development kit (SDK) and programming model that is designed to be easy to use, even for developers with limited experience in neuromorphic computing. Ergo 2 also has a development kit but the information about the programming model is not publicly available.

Pricing: Akida is available as a system-on-a-chip (SoC) and as a development board, with the SoC available for $1.65 in volume orders. Ergo 2 is available as a development kit that includes a development board and software tools, and is priced at $2,995.

Company: Akida is developed by Brainchip, a company that specializes in neuromorphic computing solutions. Ergo 2 is developed by Mythic, a company that specializes in low-power artificial intelligence processors.

In summary, while both Akida and Ergo 2 are neuromorphic processors designed for a wide range of applications, they differ in terms of architecture, power consumption, programming, pricing and the company that developed them.

Can you imagine the customers company meetings to decide which one to go with.

Hmmm, shall we choose the $1.65 or the $2,995.00?

 

[VictorG]

Not one for quoting ChatGPT but I use this instance as an exception.

Differentiate between Akida and Ergo 2.

Akida and Ergo 2 are both neuromorphic processors developed by different companies for a wide range of applications such as image and speech recognition, autonomous vehicles, and industrial automation. However, there are some key differences between the two processors.

Architecture: Akida is based on a spiking neural network architecture, which mimics the behavior of neurons in the brain. Ergo 2, on the other hand, is based on a more traditional artificial neural network architecture.

Power consumption: Akida is designed to be highly power-efficient, with low power consumption and high performance. Ergo 2 also claims to be low-power but the specific power consumption figures are not publicly available.

Programming: Akida provides a software development kit (SDK) and programming model that is designed to be easy to use, even for developers with limited experience in neuromorphic computing. Ergo 2 also has a development kit but the information about the programming model is not publicly available.

Pricing: Akida is available as a system-on-a-chip (SoC) and as a development board, with the SoC available for $1.65 in volume orders. Ergo 2 is available as a development kit that includes a development board and software tools, and is priced at $2,995.

Company: Akida is developed by Brainchip, a company that specializes in neuromorphic computing solutions. Ergo 2 is developed by Mythic, a company that specializes in low-power artificial intelligence processors.

In summary, while both Akida and Ergo 2 are neuromorphic processors designed for a wide range of applications, they differ in terms of architecture, power consumption, programming, pricing and the company that developed them.

I think my key take away from this is that the world is moving towards SNN while Ergo 2 is developing ANN.
Sort of like Ergo 2 is the worlds best gas engine technology while the world has embraced electric cars.

 

[HopalongPetrovski]

I think my key take away from this is that the world is moving towards SNN while Ergo 2 is developing ANN.
Sort of like Ergo 2 is the worlds best gas engine technology while the world has embraced electric cars.

They're making the best damn buggy whips money can buy.

 

[Dhm]

Not sure what to expect with the next 4C. Many here are wary and with good reason. If the numbers are less than expected I am hoping for some strong spin with the commentary. Last time we got most sales activity and engagement ever, or words to that effect. That is a very strong statement and we can take hope from it. But not necessarily expectation.

But the Qualcomm video could imply that they have already generated revenue for us. And if so, how many others may be doing the same in their markets? Up until a few weeks ago we hadn’t heard of VVDN then out of the blue there they are.

Interesting times ahead, especially with so many shorts that may need covering.