BRN Discussion Ongoing

[Bravo]

From the crapper…. Why is Brainchip recently a Nr. 1 topic with AI created stuff?

Nuclear Detection in the Palm of Your Hand - NeuromorphicCore.ai | NeuromorphicCore.AI

𝗡𝘂𝗰𝗹𝗲𝗮𝗿 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗶𝗻 𝘁𝗵𝗲 𝗣𝗮𝗹𝗺 𝗼𝗳 𝗬𝗼𝘂𝗿 𝗛𝗮𝗻𝗱 ☢️🧠 How researchers at Sandia National Laboratories used brain-inspired chips to identify radioactive isotopes using less energy than a single blink of a status LED. 🔬 The Breakthrough Handheld radiation detectors often run into what engineers call...

www.linkedin.com

Nuclear Detection in the Palm of Your Hand - NeuromorphicCore.ai

See how Sandia National Laboratories used Intel Loihi 2 and BrainChip Akida to identify nuclear isotopes using 70,000x less energy than conventional processors—proving that brain-inspired "spiking" algorithms like LCA outperform standard AI in messy, real-world security environments.

neuromorphiccore.ai

View attachment 96617

Re the nuclear-detection paper comparing Akida, SpiNNaker and Loihi 2 there were comments that Akida delivered the fastest inference but lower accuracy due to the 4-bit quantisation constraint.

This raises the question which generation of Akida was actually tested?

I imagine most academic work still uses the first widely available Akida hardware (AKD 1000), which typically runs models in INT4 precision to maximise power efficiency. As far as I understand it, it's great for speed and edge deployment, but a lower precision naturally introduces quantisation error that can impact accuracy.

What’s interesting is that later iterations of the Akida architecture support higher precision and mixed-precision approaches (e.g. INT8). Presumably moving from 4-bit to 8-bit precision could entail a significant recovery of accuracy.In other words, if the same benchmark were run using higher-precision modes or newer Akida generations, you might see Akida maintain its inference-speed advantage while narrowing or (hopefully) even eliminating the accuracy gap with Loihi 2??

If so, it would be very interesting to see how the results change under those conditions.

Nuclear Detection in the Palm of Your Hand - NeuromorphicCore.ai

See how Sandia National Laboratories used Intel Loihi 2 and BrainChip Akida to identify nuclear isotopes using 70,000x less energy than conventional processors—proving that brain-inspired "spiking" algorithms like LCA outperform standard AI in messy, real-world security environments.

neuromorphiccore.ai

 

[Bravo]

Do we have Renesas still on the books, what about Megachips

Screenshot of current Integration Partner page on the website.

 

[7For7]

Re the nuclear-detection paper comparing Akida, SpiNNaker and Loihi 2 there were comments that Akida delivered the fastest inference but lower accuracy due to the 4-bit quantisation constraint.

This raises the question which generation of Akida was actually tested?

I imagine most academic work still uses the first widely available Akida hardware (AKD 1000), which typically runs models in INT4 precision to maximise power efficiency. As far as I understand it, it's great for speed and edge deployment, but a lower precision naturally introduces quantisation error that can impact accuracy.

What’s interesting is that later iterations of the Akida architecture support higher precision and mixed-precision approaches (e.g. INT8). Presumably moving from 4-bit to 8-bit precision could entail a significant recovery of accuracy.In other words, if the same benchmark were run using higher-precision modes or newer Akida generations, you might see Akida maintain its inference-speed advantage while narrowing or (hopefully) even eliminating the accuracy gap with Loihi 2??

If so, it would be very interesting to see how the results change under those conditions.

Nuclear Detection in the Palm of Your Hand - NeuromorphicCore.ai

See how Sandia National Laboratories used Intel Loihi 2 and BrainChip Akida to identify nuclear isotopes using 70,000x less energy than conventional processors—proving that brain-inspired "spiking" algorithms like LCA outperform standard AI in messy, real-world security environments.

neuromorphiccore.ai

View attachment 96679

On the end it’s a waiting game… as usual. Let’s see what we will see from now on!

 

[Diogenese]

Re the nuclear-detection paper comparing Akida, SpiNNaker and Loihi 2 there were comments that Akida delivered the fastest inference but lower accuracy due to the 4-bit quantisation constraint.

This raises the question which generation of Akida was actually tested?

I imagine most academic work still uses the first widely available Akida hardware (AKD 1000), which typically runs models in INT4 precision to maximise power efficiency. As far as I understand it, it's great for speed and edge deployment, but a lower precision naturally introduces quantisation error that can impact accuracy.

What’s interesting is that later iterations of the Akida architecture support higher precision and mixed-precision approaches (e.g. INT8). Presumably moving from 4-bit to 8-bit precision could entail a significant recovery of accuracy.In other words, if the same benchmark were run using higher-precision modes or newer Akida generations, you might see Akida maintain its inference-speed advantage while narrowing or (hopefully) even eliminating the accuracy gap with Loihi 2??

If so, it would be very interesting to see how the results change under those conditions.

Nuclear Detection in the Palm of Your Hand - NeuromorphicCore.ai

See how Sandia National Laboratories used Intel Loihi 2 and BrainChip Akida to identify nuclear isotopes using 70,000x less energy than conventional processors—proving that brain-inspired "spiking" algorithms like LCA outperform standard AI in messy, real-world security environments.

neuromorphiccore.ai

View attachment 96679

Hi Bravo,

According to the Diogenese theory of everything, Akida 1000 (4-bit) may not be the ideal tool where mathematical precision in the output is required, but that does not necessarily detract from its utility in use cases where the benchmark is probability, such as video, speech, ...

4-bit allows 16 gradations, which means individual levels are ~ 6% apart, while 8-bit has 256. giving about 0.4% differentiation.

16-bit provides about 65k "levels" which is 1/256 th of 8-bit, (less than 1/500 th of 1%).

So basically, mathematical accuracy better than 6% error is difficult to achieve with Akida 1000, but where probability is concerned, that may not be such a hurdle.

Loihi 2 provides graduated quantization up to 32-bit:

https://www.intel.com/content/www/u...rphic-computing-loihi-2-technology-brief.html

It is also noteworthy that Intel 4 is in 4 nm while Akida 1000 is 28 nm. Comparing Akida 1000 with Loihi 2 does not do justice to Akida's tech advancements, but to have compared the Akida 2 on-line FPGE would also have underplayed the capabilities of the soon-to-be Akida 2 ASIC.

 

[Pom down under]

Do we have Renesas still on the books, what about Megachips

What you on about shitstick?

 

[Diogenese]

Hi Bravo,

According to the Diogenese theory of everything, Akida 1000 (4-bit) may not be the ideal tool where mathematical precision in the output is required, but that does not necessarily detract from its utility in use cases where the benchmark is probability, such as video, speech, ...

4-bit allows 16 gradations, which means individual levels are ~ 6% apart, while 8-bit has 256. giving about 0.4% differentiation.

16-bit provides about 65k "levels" which is 1/256 th of 8-bit, (less than 1/500 th of 1%).

So basically, mathematical accuracy better than 6% error is difficult to achieve with Akida 1000, but where probability is concerned, that may not be such a hurdle.

Loihi 2 provides graduated quantization up to 32-bit:

https://www.intel.com/content/www/u...rphic-computing-loihi-2-technology-brief.html

View attachment 96682
View attachment 96683

It is also noteworthy that Intel 4 is in 4 nm while Akida 1000 is 28 nm. Comparing Akida 1000 with Loihi 2 does not do justice to Akida's tech advancements, but to have compared the Akida 2 on-line FPGE would also have underplayed the capabilities of the soon-to-be Akida 2 ASIC.

Given that Loihi 2 has uo to 32 bits, that means that it takes up a lot more wafer real estate than Akida 2 at the same architecture size (eg, 4 nm). This makes it much more expensive to make. I don't know the yield (working chips per wafer) but it tends to decrease with size, so, even though it is possible to fit more chips per wafer, the number that don't work will probably be higher than with a larger architecture.

 

[Jacob Kiraz]

What you on about shitstick?

Witz 123 counting the beat 7,8,9

 

[AusEire]

Witz 123 counting the beat 7,8,9

How much rent do you charge Wilzy for living in your head?

 

[Jacob Kiraz]

How much rent do you charge Wilzy for living in your head?

Why are you worried its between me and him

 

[Diogenese]

Given that Loihi 2 has uo to 32 bits, that means that it takes up a lot more wafer real estate than Akida 2 at the same architecture size (eg, 4 nm). This makes it much more expensive to make. I don't know the yield (working chips per wafer) but it tends to decrease with size, so, even though it is possible to fit more chips per wafer, the number that don't work will probably be higher than with a larger architecture.

... not to mention TENNs ...

... and solid state switching to replace packet switching ...

 

[Esq.111]

Given that Loihi 2 has uo to 32 bits, that means that it takes up a lot more wafer real estate than Akida 2 at the same architecture size (eg, 4 nm). This makes it much more expensive to make. I don't know the yield (working chips per wafer) but it tends to decrease with size, so, even though it is possible to fit more chips per wafer, the number that don't work will probably be higher than with a larger architecture.

Good evening Diogenese ,

Interesting ones thoughts , thankyou as always.

Personally, is still truely bewildered.

WHAT THE FARK , a device to pick up elements in the atomic , seriously could damage life as one knows it ..... yet thay are STILL FUCKIN AROUND.

TRUELY BEWILDERD BY THE LACKSADASEYH ATTITULDE DISPLAYED BY THOSE THAT HAVE THE BUDGETS YET STILL ARE FUCKING AROUND AIMLESSLY.

In the mean time ..... apparently it's fine to spend many US $ Billions to locate & secure 400~500 Kg of atomic material in Iran.

One could only come to the conclusion , those in controle .... fark the people , solong as a few get to prosperous greatly , WHO CARES.

VERY , VERY SADE STATE OF AFFAIRS.

THE TECK IS THERE , WAITING , FUCKING USE IT.

Obviously not directed at you.


FAaaark Me Spinning.


Regards,
Esq.

 

[Diogenese]

As Count Oxenstierna said to his son "You do not know, my son, with how very little wisdom the world is governed"

Of course, that was before the interweb, ...

 

[AusEire]

Why are you worried its between me and him

Does he know that he's living in your head?

I'm worried because Wilzy hasn't posted on here in over 12months. Have you considered therapy?

 

[manny100]

... not to mention TENNs ...

... and solid state switching to replace packet switching ...

If there was anything better than AKIDA1000 out there then i suggest Kevin Johnson, IBM would be using it for his system builds.
IMO these builds are warm ups for him. His builds will likely become more complex by the time he gets to AKD 2500/TENNs.

 

[Townyj]

Does he know that he's living in your head?

I'm worried because Wilzy hasn't posted on here in over 12months. Have you considered therapy?

Not sure why he thinks Pom is Wilzy in any online shape or form...

I have linked some therapy support links below, as you may be right and this person needs to seek help.

Online therapy (eTherapy)

eTherapy, online counselling or therapy refers to psychological support, therapy, education and assistance via the internet or an online platform.

www.healthdirect.gov.au

Australia's #1 Online Therapy Platform | Talked

Book affordable therapy sessions with qualified Australian therapists and psychologists. See verified therapist reviews and claim your Medicare rebate.

www.talked.com.au

 

[Space Cadet]

Pop me down as c) too

Me too. I asked the missus and she confirmed in no uncertain terms

SC

 

[Space Cadet]

Does he know that he's living in your head?

I'm worried because Wilzy hasn't posted on here in over 12months. Have you considered therapy?

Or an exorcism

SC

 

[AusEire]

Or an exorcism

SC