BRN Discussion Ongoing

[stuart888]

Interesting that Edge Impulse posted the following on LinkedIn, Akida maybe?

Edge Impulse on LinkedIn: John Deere's New Robot Planter Could Slash Fertilizer Use by 60%

John Deere's new ExactShot robotic technology relies on sensors and machine learning to reduce fertilizer use by as much as 60%, saving farmers money and…

www.linkedin.com

Perhaps they (John Deere) are using the Edge Impulse toolset?

They have a great objective:

 

[Fact Finder]

I scan the QR code and got the link to the above demo video.
View attachment 26383

Learning 🏖

They stole Rob Telson’s fan.

 

[Foxdog]

Innnnteresting

TDK Corporation (TSE:6762) today announced the TDK i3 Micro Module, the world’s first module with built-in edge AI and wireless mesh connectivity capability.

View attachment 26375
"Because it is an ultracompact, battery-powered wireless sensor module, users can achieve sensing at any desired position without physical constraints like wiring."

"By attaching this compact module to equipment, various sensors monitor conditions like vibration, temperature and sound in real time. A variety of anomalies and failures can be detected in a timely fashion."

"Data detected by the sensors are processed by i3 Micro Module’s embedded edge AI. The data do not need to be aggregated and analyzed in the cloud, minimizing network traffic. Modules connect to each other through a wireless mesh network, which is formed automatically between modules simply by installing them."

https://www.tdk.com/en/news_center/press/20230105_01.html

https://www.tdk.com/en/featured_stories/entry_046-sensor-micro-module.html

Better have AKIDA in it, otherwise we have direct competition?

 

[Kachoo]

Edge Impulse on LinkedIn: Syntiant to Introduce Turnkey Edge AI Security Solution at CES 2023

Syntiant Corp. is showing off its edge AI audio, sensor and vision processing capabilities with several turnkey solutions at #CES2023 — including the…

www.linkedin.com

Just posting some edge impulse items that they have marketed at the CES shows how intertwined

Pretty sure we didn't need to see this a third time (first @Gman mislead the forum by claiming this was BRN... then you twice...)

Deleted mine I did not see that post.

 

[wilzy123]

Better have AKIDA in it, otherwise we have direct competition?

Not necessarily direct competition, since I cannot vouch for the capability of the thing doing the AI in that device... merely shared, because it does sound like something Akida is particularly good at.

 

[Realinfo]

The dominoes are beginning to fall.

After the ‘Eyes‘ were unable to find a true blue competitor, it was always only a matter of time before the likes of Intel swallowed their pride and joined the Akida revolution. They just had to find a way to save face before doing so, and inviting Brainchip to join their IFS programme was the perfect foil. Have no doubt, a license is not far away.

Can you imagine the panic going on within the hierarchy of IBM and other biggies right now. If they want to stay in the game, they must ”do an Intel” and join the revolution...they simply have no choice.

If they still think they can take another route, the imminent release of Akida 2.0 will blow these thoughts asunder. Everyone will eventually use Akida in some way. It’ll be in everything…even your Weet-Bix !!

The thought of using NDA’s and other confidentiality agreements amuses me, cos they’ve now clearly passed their use by date. They’re simply mechanisms where everyone is trying to keep the same secret from everybody else…AKIDA !!

Those who harness the beast best…win !

In the final paragraph above, I suggest that it’s no longer Akida, or if you’re using it that must remain a secret. Like the colt from Old Regret, Akida has got away. It hasn’t just escaped…it’s bolted into the marketplace. If anyone either directly or indirectly associated with the tech world still hasn’t heard of it by now, they must be living under a rock.

The secret that must now be kept under absolute wraps until your product at least hits the market, is how you’re applying the science fiction that is Akida.

It’s no longer a question of if you’re using it…but how.

This is why companies at CES are only talking in big picture language, devoid of any details or specifics about how they are gunna achieve what they’re talking about. Even Mercedes realise, that to reveal anymore than what they said last year, is giving their competitors too much of a heads-up on what they’re doing.

Everybody will eventually use Akida in some way or another. Those first to market will certainly gain an advantage, but it’s those who apply it’s ‘secret sauce’ in the most innovative, imaginative, ingenious, inventive way, who’ll be the big winners over their competitors.

’If you can fill the unforgiving minute with sixty seconds of brilliance…
yours is the earth and everything in it’...Rudyard Kipling.

 

[SharesForBrekky]

This Stockhead article talks about their tech trends for 2023 and the ASX listed companies that could have an impact. Brainchip gets a good mention:

https://www.theaustralian.com.au/business/stockhead/seven-tech-trends-and-asx-stocks-ready-to-get-a-taste-of-action/news-story/37712dad99a76249f098dffa3856f102

"BrainChip Holdings (ASX:BRN) is involved in neuromorphic computing, a type of AI that simulates the functionality of the human neuron, and is working on commercialisation of its Akida chip.

Akida basically mimics the human brain to analyse only essential sensor inputs at the point of acquisition, processing data with efficiency, precision, and economy of energy.

BRN says keeping machine learning local to the chip, independent of the cloud, also dramatically reduces latency while improving privacy and data security.

This quarter the company is focusing on key sales targets and converting technical evaluations into paid licenses.

The company has a market cap of $1b and cash balance as of September 30, 2022, of $US24.6m."

 

[buena suerte :-)]

Could we possibly be involved??​

By 9News Staff
7:41am Jan 7, 2023

How your phone may soon be able to tell if you're over the limit​

Recording a quick voice memo could soon tell you whether you're over the limit.
Melbourne researchers are using artificial intelligence to analyse audio waves of people who've had too much to drink, in the hope it's an early indicator of when to slow down.
Researchers from Latrobe University are currently working on something faster than a blood test or a breathalyser.

Aussie researchers from Melbourne's Latrobe University are testing artificial intelligence analysis of voice memos as a way of working out if people have had too much to drink, hoping to turn it into an app. (9News)
More than 12,000 audio recordings from Germany of sober and less sober people were input into an algorithm with artificial intelligence then trying to distinguish whether the voices are over the legal limit or not.
At the moment it's about 68 per cent accurate but the more data, the smarter the AI becomes.


"We are on a good way it's a kind of proof of concept that it's possible," Professor Emmanuel Kuntsche told 9News.

More than 12,000 audio recordings from Germany of sober and less sober people were input into an algorithm. (9News)
The next step is for the tech to be tested with Aussie voices.
Researchers say several years down the track, if it works, it could once day be put into an app to give people an early indicator of their intoxication levels before they make decisions like getting behind the wheel.
For those committed to reducing the number of drunk drivers on Australian roads, it's a move in the right direction.

The next step is for the tech to be tested with Aussie voices. (9News)
"Whether it's technology or education or behaviour change, if it can save one life, absolutely," Annie Trainor, Director of Drink Drive Victoria told told 9News.
"As long as people are getting the message, it doesn't matter where it comes from."

 

[VictorG]

Some perspective,
2021 Wanca's - What's a neuromorphic chip anyway.
2022 Brainchip - ASX 200, Many new partnerships, on everyone's lips.
2023 Akida - Starting to emerge in everything, everywhere.
2024 $$$$$$$$

 

[Bobbyk4819]

https://www.tdk.com/en/news_center/press/20230105_01.html

Partnering with Texas Instruments

BrainChip Appoints Duy-Loan Le to its Board of Directors

LAGUNA HILLS, CA / ACCESSWIRE / October 31, 2022 / BrainChip Holdings Ltd (ASX:BRN)(OTCQX:BRCHF)(ADR:BCHPY), the world's first commercial producer of ultra-low power neuromorphic AI IP, announces the appointment of Duy-Loan Le to its Board of Directors, as a non-executive director. Mrs. Le has a...

www.accesswire.com

"having retired from Texas Instruments as a Senior Fellow after 35 years. While at TI, she led TI's multi-billion-dollar memory and DSP product lines with joint venture partners in five countries and three continents"

let's hope Duy Loan Le has been involved..

 

[Diogenese]

I am so proud!

First time round, I missed the face detection on the right.

Face, taste, vibration ...
In cabin detection?

 

[jtardif999]

The other day someone suggested Brainchip only had one product.

I think this is an inappropriate way to look at AKIDA technology.

AKIDA technology is at its heart a neural processing engine (NPE) AKD1000/AKIDA 1.0 whatever name you want to give it is made up of 80 of these Neural Processing Units (NPUs).

AKD1000 is designed to allow for 64 x AKD1000 to be ganged together and this would then be a semiconductor of 5,120 NPU’s

Brainchip has said they could if needed gang 1,024 AKD1000 chips together making a total of 81,920 NPU’s.

Renesas licenced just two nodes or 8 NPE’s.

Anil Mankar stated at the 2021 Ai Field Day that you could do keyword spotting with just one node if you did not have latency concerns but 1.5 to 2 nodes would likely be desirable and referenced Renesas.

MegaChips licensed an 80 NPU AKD1000 IP.

Edge Impulse is comparing AKD1000 with a GPU.

So my point is Brainchip does not have just one semiconductor it has on its IP shelf 81,920 variants of that semiconductor IP which can be sold for an ultra low powered single purpose MCU right through to a computing task requiring the computing power of 81,920 NPUs or 1,024 AKD1000 chips.

1,024 AKD1000 chips would give you the computing power of 1,024 GPUs for a total cost of $25,600 being 1,024 x $25.00.

As Peter van der Made said in his Pitt Street Research interview in 2021 ‘we are only just getting started.’

Something further occurred to me which if you doubt my proposition may convince you.

If you look up Intel’s Loihi 1 & 2 you will see the actual manifestation of what I am alluding to as they start with two Loihi chips to make their base product USB then gang together in different numbers to create further variants of Loihi with different Hawaiian names right up to their attempt at a full blown neuromorphic server and claim each one as a separate distinct product.

My opinion only DYOR
FF

AKIDA BALLISTA

I’m going to mess up your math FF. Akida1.0 has 80 nodes and each node has 4 NPUs. Everywhere you refer to the number of NPUs can be multiplied by 4

 

[Fact Finder]

Not necessarily direct competition, since I cannot vouch for the capability of the thing doing the AI in that device... merely shared, because it does sound like something Akida is particularly good at.

Hi @wilzy

The following indicates that whatever TDK are using it is not AKIDA and it is far from as advanced in fact it is primitive in comparison:

“We also plan to further mature the embedded edge AI to enable it to continuously learn the situation at the point of installation and make autonomous decisions, thereby evolving the module to be able to support the types of advanced decision-making required at the forefront of integration.”

They are planning to have a plan without a timeframe.

They are silent on battery life. They also say they are in the throws of improving the battery which is a strange thing to say as you release a product.

I think we should hold close what the CEO Sean Hehir said at the AGM about claims being made that do not come near the AKIDA science fiction.

As you know it was Edge Impulse who used Science Fiction to describe AKIDA technology not Sean Hehir but I just like saying it because it is well ‘SCIENCE FICTION.’

Anyway that’s the last time I will use Science Fiction to describe AKIDA technology until the next time I use it.

Did I mention how being Science Fiction makes the Patents stronger and easier to defend.

You just have to love being SCIENCE FICTION.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Boab]

Those who harness the beast best…win !

In the final paragraph above, I suggest that it’s no longer Akida, or if you’re using it that must remain a secret. Like the colt from Old Regret, Akida has got away. It hasn’t just escaped…it’s bolted into the marketplace. If anyone either directly or indirectly associated with the tech world still hasn’t heard of it by now, they must be living under a rock.

The secret that must now be kept under absolute wraps until your product at least hits the market, is how you’re applying the science fiction that is Akida.

It’s no longer a question of if you’re using it…but how.

This is why companies at CES are only talking in big picture language, devoid of any details or specifics about how they are gunna achieve what they’re talking about. Even Mercedes realise, that to reveal anymore than what they said last year, is giving their competitors too much of a heads-up on what they’re doing.

Everybody will eventually use Akida in some way or another. Those first to market will certainly gain an advantage, but it’s those who apply it’s ‘secret sauce’ in the most innovative, imaginative, ingenious, inventive way, who’ll be the big winners over their competitors.

’If you can fill the unforgiving minute with sixty seconds of brilliance…
yours is the earth and everything in it’...Rudyard Kipling.

Your enthusiasm is infectious. Love it.

 

[Fact Finder]

I’m going to mess up your math FF. Akida1.0 has 80 nodes and each node has 4 NPUs. Everywhere you refer to the number of NPUs can be multiplied by 4

So you think I am a down ramper.

 

[wilzy123]

They are silent on battery life. They also say they are in the throws of improving the battery which is a strange thing to say as you release a product.

Absolutely, which is why it pays to scrutinise these things (as you have).

 

[VictorG]

And another dot to put on the radar

Krishnamurthy Vemuru on LinkedIn: Implementation of the Canny Edge Detector Using a Spiking Neural Network

I am happy to share my latest publication on the Neuromorphic version of Canny Edge Detector implemented with a Spiking Neural Network of conductance based…

www.linkedin.com

 

[VictorG]

Hi @wilzy

The following indicates that whatever TDK are using it is not AKIDA and it is far from as advanced in fact it is primitive in comparison:

“We also plan to further mature the embedded edge AI to enable it to continuously learn the situation at the point of installation and make autonomous decisions, thereby evolving the module to be able to support the types of advanced decision-making required at the forefront of integration.”

They are planning to have a plan without a timeframe.

They are silent on battery life. They also say they are in the throws of improving the battery which is a strange thing to say as you release a product.

I think we should hold close what the CEO Sean Hehir said at the AGM about claims being made that do not come near the AKIDA science fiction.

As you know it was Edge Impulse who used Science Fiction to describe AKIDA technology not Sean Hehir but I just like saying it because it is well ‘SCIENCE FICTION.’

Anyway that’s the last time I will use Science Fiction to describe AKIDA technology until the next time I use it.

Did I mention how being Science Fiction makes the Patents stronger and easier to defend.

You just have to love being SCIENCE FICTION.

My opinion only DYOR
FF

AKIDA BALLISTA

Enough with the Science Fiction, we need more cowbell

 

[Boab]

Why Luminar's CEO is running over fake kids in a Tesla at CES

Austin Russell is trying to prove that automotive-safety systems with lidar — the technology at the core of his company — are safer than those without.

www.marketwatch.com

 

[Tothemoon24]

CES - Health Ai / Booth 24

Sunburn suffering .

In search of a good nights sleep .
Take 1x 50mg viagra approximately 1hr before bed.

Please note : this will not cure the sunburn but it will keep the sheets off your skin .