BRN Discussion Ongoing

[Lex555]

Hi JK,

Akida's forte is identifying (classifying) input signals from sensors. In simple applications this may be sufficient to trigger a direct response or action.

However, in some cases the Akida output is used as an input to another CPU/GPU (von Neumann processor) to form part of that computer's program variables.

In the first case, the entire process gets the full power saving/speed improvement from Akida.

In the second case, the benefit is the reduction in power/time which Akida brings to the classification task while the CPU performs the remaining processes under the control of its software program. This is important because the classification task carried out on a software controlled CPU uses very large amounts of power and takes a relatively long time.

Classification of an image on a CPU uses CNN (convolutional neural network) processes which involve multiplying multi-bit (8, 16, 32, 63) bytes representing each pixel on the sensor. Multiplication involves the number of computer operations determined by the square of the number of bits in the byte, so an 8-bit byte multiplication would involve, 64 computer operations. For 32-bit bytes, 1024 operations are required to process the output from a single pixel, whether it's value has changed or not.

On the other hand, Akida ignores pixels whose output value does not change, and only performs a computer operation for the pixels whose output changes (an event). This is "sparsity". In addition, in 1-bit mode there is only a singe computer operation for each pixel event.

For example, the sparsity may reduce the number of events by, say, 40%.

Even in 4-bit mode, Akida only needs 16 computer operations, and that only for pixels whose output has changed.

Hence there are large savings in power and time in using Akida to do the classification task compared to using, eg, a 32-bit ARM Cortex microprocessor.

While the rest of the program may be carried out on the microprocessor, this uses comparatively little power compared to the power the microprocessor would have used performing the CNN task. So there are still large power savings to be made by using Akida in "accelerator" mode as an input device for a von Neumann processor.

The other point is that Akida performs its classification independent of any processor with which it is associated. For example, Akida 1000 includes an ARM Cortex processor, but this is only used for configuration of the arrangement of the NPU nodes to optimize performance for the particular task, but the ARM Cortex plays no part in actual classification task. The ARM Cortex does not form part of the Akida IP. Akida is "processor-agnostic" and can operate with any CPU.

Diogenese this is a great summary in layman’s terms that I was looking for. @zeeb0t i think this is worthy of a sticky at top of BRN for all new readers to cast an eye over, maybe like a tech 101 thread

 

[hotty4040]

all the talk of birds I have to show you the bird which came into my garden today.

The real crunch is in the following:

IN BRIEF​ ASX recently advised that “ramping” announcements are on its radar and that companies making them risk suspension from the ASX. “Ramping” announcements are designed to increase a company’s share price, rather than to inform the market. ASX has observed that these types of announcements are often disclosures about customer contracts. Companies should be aware that ASX is maintaining its spotlight on the disclosure of customer contracts, following its related guidance issued late last year on this topic. So a company such as Brainchip takes the risk releases on the ASX then retrospectively has the ASX rule on their notice and decide that it is ramping and suspends them. The reason to take the risk being some short term price improvement with a downside of suspension from trading. If they put it up as non price sensitive the MF’s of this world can broadcast to all the world not price sensitive no money in this it is another fluff announcement. Loose, loose, loose every way you look at it. The downside of suspension from trading outweighs a non price sensitive ASX release. I can guarantee that any lawyer would faced with the risk reward involved advise every client not release on ASX make it a press release no downside. Release on the ASX and you could be suspended and shareholders could sue you for negligence. All this lay opinion takes us nowhere. Tony Dawe and the two Australian Directors have recently had a meeting with the head of the ASX trying to get guidance and the response was if you were non compliant we would tell you and until we tell you otherwise you will remain so. No help just we will punish you if you get it wrong. So who here wants the company to risk suspension with their fingers crossed for a short term price movement when a press release is no risk and just like Mercedes Benz non price sensitive announcement can drive the price if the market likes the news. Most ordinary retail have never heard of ARM. The tech savvy will be reading ARM’s Twitter and the technology mags and blogs and understand. My opinion only DYOR FF AKIDA BALLISTA

ARMe's for the poor, seems to fit, somehow here, maybe not.


Akida Ballista

ARMen to that.

My opinion only DYOR

hotty...

 

[Hrdwk]

Hi @Zedjack33
This is a great find generously shared and I love the WOW but only 3 people have said WOW apart from me.

EVERY ONE NEEDS TO OPEN THE LINK AND WATCH A DRONE BEING VOICE CONTROLLED VIA AKIDA.

It makes very relevant Zedjack’s further posts.

My definite opinion DYOR
FF

AKIDA BALLISTA

What would be totally mind blowing would be for them to do a Duracell battery test to show one drone with Akida and one without and the difference in flight time.

Can you suggest this to BRN?

 

[Slymeat]

That drone demo made me think, as it is using single word commands that could occur in everyday situations. When controlling similar devices, we will need to be attentive to their “desire” to respond.

I found the response to the “go” command to be intriguing. The drone recognised the command but seemed to imply, by its response, a “go where?“. And I thought exactly the same.

I wonder what the world will be like when we are surrounded by devices using Akida voice/gesture/eye movement detection etc. even smells. We will have to start being very careful about some of the things we do and say in our normal day activities As devices may imadvertently jump to attention and carry out our unintentional commands.

One can only imagine the unexpected things that could result from utterances and gestures made whilst, say, watching sport on TV.

 

[buena suerte :-)]

Sorry but sometimes they are a pain in the arse! Come to school with chicks stopping students walking around. (not allowed near)
Come in your yard and once even in my house! THEY ARE DANGEROUS!
Had one nearly attack my Grand Cheerokee seeing its reflection in windows and mirrors. (close one!)
One lady nearby had her BMW trashed by a Cassie as it saw its reflection in the mirrors, went one side then the other! Imagine reporting that to Insurance. Doors dented , mirrors gone.

And can they KICK! Like a Roo , just up and BOTH FEET. attacked my neighbours glass door that way.

Nice to see over the road.......but closer and start watchin out what their up to.

Yak52

Wow I didn't realise they were so aggressive as you say nice from a distance

 

[RobjHunt]

But thistle do for now
for some reason it didn’t post last night.

 

[Fact Finder]

Something to consider that might not have jumped out and grabbed you by the scruff of the neck if you are not a @Diogenese is SCALABILITY.

If you have listened to Anil Mankar at the 2021 Ai Field Day you will recall he discussed with an audience member that if you are prepared to accept some latency increase you can use just ONE node of the AKIDA neural fabric.

The AKD1000 has 80 nodes and is designed to allow you to gang up to 64 AKD1000 should you need the power of 5,120 nodes.

What you may not be aware of is that 64 was chosen from a theoretically possible 1,024 AKD1000 chips being ganged together giving you the neuromorphic processing power of 81,920 AKIDA nodes.

When @uiux and @Zeebot were having the Nviso in cabin system demonstrated to them it was stated by Tim Llewellyn that they were not using anywhere near the full processing power of their one AKD1000 chip that you see in the vehicle and they intend to add further functions via this one chip.

So what is the point well clearly AKD1000 is ridiculously powerful and this is why the former CEO Mr. Dinardo stated that 64 AKD1000 chips was chosen because it would more than cover their target market.

However the AKIDA IP can run with 1 node right up to 81,920 nodes.

In a way what this scalability means is infinite product applications in fact imagining a product that might require 81,921 nodes is beyond my imagination.

So when Brainchip refers to a competitors product as not being scalable this is a HUGE barrier to that product taking market share from Brainchip.

My opinion only so DYOR
FF

AKIDA BALLISTA

 

[Boab]

Question for somebody to answer that is bugging me:

If our sales model is now focusing on IP rather than the chip itself, how do we realise the massive power savings when the IP is at the mercy of the host chip design power le

Something to consider that might not have jumped out and grabbed you by the scruff of the neck if you are not a @Diogenese is SCALABILITY.

If you have listened to Anil Mankar at the 2021 Ai Field Day you will recall he discussed with an audience member that if you are prepared to accept some latency increase you can use just ONE node of the AKIDA neural fabric.

The AKD1000 has 80 nodes and is designed to allow you to gang up to 64 AKD1000 should you need the power of 5,120 nodes.

What you may not be aware of is that 64 was chosen from a theoretically possible 1,024 AKD1000 chips being ganged together giving you the neuromorphic processing power of 81,920 AKIDA nodes.

When @uiux and @Zeebot were having the Nviso in cabin system demonstrated to them it was stated by Tim Llewellyn that they were not using anywhere near the full processing power of their one AKD1000 chip that you see in the vehicle and they intend to add further functions via this one one chip.

So what is the point well clearly AKD1000 is ridiculously powerful and this is why the former CEO Mr. Dinardo stated that 64 AKD1000 chips was chosen because it would more than cover their target market.

However the AKIDA IP can run with 1 node right up to 81,920 nodes.

In a way what this scalability means is infinite product applications in fact imaging a product that might require 81,921 nodes is beyond my imagination.

So when Brainchip refers to a competitors product as not being scalable this is a HUGE barrier to that product taking market share from Brainchip.

My opinion only so DYOR
FF

AKIDA BALLISTA

Wow, the penny just dropped. Wow again.

 

[TheUnfairAdvantage]

Something to consider that might not have jumped out and grabbed you by the scruff of the neck if you are not a @Diogenese is SCALABILITY.

If you have listened to Anil Mankar at the 2021 Ai Field Day you will recall he discussed with an audience member that if you are prepared to accept some latency increase you can use just ONE node of the AKIDA neural fabric.

The AKD1000 has 80 nodes and is designed to allow you to gang up to 64 AKD1000 should you need the power of 5,120 nodes.

What you may not be aware of is that 64 was chosen from a theoretically possible 1,024 AKD1000 chips being ganged together giving you the neuromorphic processing power of 81,920 AKIDA nodes.

When @uiux and @Zeebot were having the Nviso in cabin system demonstrated to them it was stated by Tim Llewellyn that they were not using anywhere near the full processing power of their one AKD1000 chip that you see in the vehicle and they intend to add further functions via this one one chip.

So what is the point well clearly AKD1000 is ridiculously powerful and this is why the former CEO Mr. Dinardo stated that 64 AKD1000 chips was chosen because it would more than cover their target market.

However the AKIDA IP can run with 1 node right up to 81,920 nodes.

In a way what this scalability means is infinite product applications in fact imaging a product that might require 81,921 nodes is beyond my imagination.

So when Brainchip refers to a competitors product as not being scalable this is a HUGE barrier to that product taking market share from Brainchip.

My opinion only so DYOR
FF

AKIDA BALLISTA

I think Sean was making reference to this at the AGM. Something along the lines - we exceed where low power is required, but when power is available and not an issue Akida is also a very powerful compute solution. I think this is also the key to the words “beyond the edge”. The edge is now not the only target!

Edit: not sure if said at AGM or in the presentation with Mark afterwards.

 

[Mugen74]

Hi @Zedjack33
This is a great find generously shared and I love the WOW but only 3 people have said WOW apart from me.

EVERY ONE NEEDS TO OPEN THE LINK AND WATCH A DRONE BEING VOICE CONTROLLED VIA AKIDA.

It makes very relevant Zedjack’s further posts.

My definite opinion DYOR
FF

AKIDA BALLISTA

Lack of wows possibly because it was already posted 3 hrs earlier

 

[M_C]

Any links to Microsoft?

BRN Discussion Ongoing

I believe Brainchip's emerging partnership strategy as a pathway to IP sales is proving to be the epitome of the old saying "You don't get a dog then do the barking yourself " Arm is a very big dog. AIMHO TT

thestockexchange.com.au

 

[ndefries]

I think Sean was making reference to this at the AGM. Something along the lines - we exceed where low power is required, but when power is available and not an issue Akida is also a very powerful compute solution. I think this is also the key to the words “beyond the edge”. The edge is now not the only target!

Edit: not sure if said at AGM or in the presentation with Mark afterwards.

I think beyond the edge is a term we will keep thinking about! If i was needing to do work in outer space and needed processing for a very long time before coming back to earth i would defintely call that situation beyond the edge!

 

[alwaysgreen]

I think Sean was making reference to this at the AGM. Something along the lines - we exceed where low power is required, but when power is available and not an issue Akida is also a very powerful compute solution. I think this is also the key to the words “beyond the edge”. The edge is now not the only target!

Edit: not sure if said at AGM or in the presentation with Mark afterwards.

I agree with this interpretation of Sean's comment

 

[Shadow59]

I think beyond the edge is a term we will keep thinking about! If i was needing to do work in outer space and needed processing for a very long time before coming back to earth i would defintely call that situation beyond the edge!

I love how we're all talking about "beyond the edge" now.
That's because we've conquered the edge, so we can now go where no one has gone before.
The rest have still got to get to the edge yet.

 

[Sirod69]

BRN Discussion Ongoing

I believe Brainchip's emerging partnership strategy as a pathway to IP sales is proving to be the epitome of the old saying "You don't get a dog then do the barking yourself " Arm is a very big dog. AIMHO TT

thestockexchange.com.au

perhaps only 2 days ago and we see a link

 

[ndefries]

Wow I didn't realise they were so aggressive as you say nice from a distance

Wow got me thinking if these drones only needed mostly solar power because they were so energy efficent with Akida and had lots of back up local building charge points along with Akida vibration analysis for maintenance they could be constantly on duty on behalf of police and those that look after community assets - i.e power poles in remote locations.

programmed to head to pre set up GPS locations when out of internet connectivity and return back with content comparing the asset image to what it should be and then alerting and requesting maintenance and parts. The time and labour cost saving is pretty noticable.

The use cases are pretty huge!

 

[Boab]

I love how we're all talking about "beyond the edge" now.
That's because we've conquered the edge, so we can now go where no one has gone before.
The rest have still got to get to the edge yet.

You all know what he says.........

 

[Sotherbys01]

Beyond The Edge - Official Trailer

Available now on demand and DVD!A decade into their mission to reach the end of the universe, Marine biologist Dr. Abe Anderson (Sean Maher) and pilot Lt. Co...

www.google.com

 

[Fact Finder]

A bit of light reading but with a positive message for companies like Ford and Mercedes where Brainchip is concerned:

MSN

www.msn.com

My opinion only DYOR
FF

AKIDA BALLISTA

 

[MDhere]

You all know what he says.........
View attachment 7756

Matter of fact I have an infinity tatt too

But I also like this little edge dude