BRN Discussion Ongoing

Diogenese

Top 20
Just back to Valeo.

Posted this about this time last year with some of the authors being Valeo.

Were playing with CNN2SNN simulations.

Full paper in original post & I added the conclusion snip this time.



Don't know if been posted already, but skimming some posts around Valeo earlier and this paper released recently, June 6th.

Paper attached.

At least we have Valeos name for it...Spiking Lidar......SpikiLi :D

Like to think they just replicating / simulating what Akida can already do for their own confirmations as part of the process.

No direct reference to us though :cautious:

Maybe @Diogenese will understand the implications better and explain in simpler terms for us.

IMG_20220614_215425.jpg



IMG_20220614_220217.jpg


View attachment 40027
Talk about morphic resonance!

In my conversation with @Proga last night, I suggested that the milestone payments under the Valeo JD agreement could be attached to improving accuracy and converting Valeo's CNN model library to SNN, and this could have been the motivation for the "customer-suggested" changes to the Akida engineering sample SoC which produced Akida 1 with CNN2SNN and 4-bit weights/activations.

My crystal ball is beside itself with joy.
 
  • Like
  • Love
  • Fire
Reactions: 73 users

IloveLamp

Top 20
  • Like
  • Fire
Reactions: 9 users

Proga

Regular
Talk about morphic resonance!

In my conversation with @Proga last night, I suggested that the milestone payments under the Valeo JD agreement could be attached to improving accuracy and converting Valeo's CNN model library to SNN, and this could have been the motivation for the "customer-suggested" changes to the Akida engineering sample SoC which produced Akida 1 with CNN2SNN and 4-bit weights/activations.

My crystal ball is beside itself with joy.
Indeed 💪
 
  • Like
  • Fire
Reactions: 4 users
Just back to Valeo.

Posted this about this time last year with some of the authors being Valeo.

Were playing with CNN2SNN simulations.

Full paper in original post & I added the conclusion snip this time.



Don't know if been posted already, but skimming some posts around Valeo earlier and this paper released recently, June 6th.

Paper attached.

At least we have Valeos name for it...Spiking Lidar......SpikiLi :D

Like to think they just replicating / simulating what Akida can already do for their own confirmations as part of the process.

No direct reference to us though :cautious:

Maybe @Diogenese will understand the implications better and explain in simpler terms for us.

IMG_20220614_215425.jpg



IMG_20220614_220217.jpg


View attachment 40027


Thanks for posting that @Fullmoonfever.

I noticed one of the authors came from CEA-List and couldn’t recall them. After looking at their web page they look pretty important in the scheme of things as I understand they provide the critical function of certifying the AI to enable its use. We might have them to thank for validating Akida‘s use for Valeo. The date looks right, Valeo signed an agreement with us in June 2020 and in September CEA-List were involved in formal validation which might have been one of the milestones mentioned in the agreement.


1689392393255.png


1689392509823.png




A few other companies on their books:


1689392239908.png


:)
 
  • Like
  • Love
  • Fire
Reactions: 28 users

overpup

Regular
Let me settle this one as well!
This was only a couple of weeks ago!!!
Now can everybody just accept that Mercedes and Brainchip are heavily involved with each other and that they see akida as 5-10x more efficient than anything else out there…..

View attachment 40026
I love how his reply was kind of blunt: Like you would have to be a complete moron to miss the connections between MB and BrainChip :rolleyes:
 
Last edited:
  • Haha
  • Like
  • Fire
Reactions: 22 users

Proga

Regular
I love how his reply was kind of blunt: Like you would have to be a complete moron to miss the connections between MD and BrainChip :rolleyes:
...Progo I am looking in your direction
🤣sounds like one of those dumb comments people make when joining a conversation halfway through

I wasn't buying a current connection between MB and Brainchip on the basis MB hadn't deleted an article on their website about the EQXX.
 
Last edited:
  • Like
  • Love
Reactions: 6 users

mrgds

Regular
Hey Chippers

As some will know, im very interested/invested in Tesla along with Brainchip.

Very recently another Musk led company has emerged called xAI,
A very respected "you tuber " Rob Maurer, covered the 1st Twitter post on xAI, and something i thought very obvious, yet exciting
is a couple of the bullet points Musk and his xAI team stated ,

* extreme silicon shortage today to voltage transformer shortage in a year, then electricity shortage in another year.
* MOST IMPORTANT METRIC OVER TIME: USEFUL COMPUTE PER UNIT OF ENERGY

then

* If i look at the experience with Tesla, what weve discovered over time is that weve actually overcomplicated the problem.
I can"t speak in too much detail about what Tesla"s figured out, except to say that, in broad terms, the answer was much simplier than we thought. We were too dumb to realize how simple the answer was.
* IN THE CONTEXT OF "AGI" BEING BRUTE FORCED


So, we all know of the energy efficiency of Akida, ....................... So, it begs the question of, Is this the implementation of SNN"s ?

Link below,


Thats my hopeful dot for the weekend, BTW .............. thanks to all contributers re the Valeo/Scala3 .......... CNN/SNN ................ "YOU ALL ROCK "

AKIDA BALLISTA .................... or maybe "AKIDESLA TESLLISTA"
 
  • Like
  • Love
  • Fire
Reactions: 52 users

Sirod69

bavarian girl ;-)
Haven´t heard it yet, but afternoon I will have time for it.
https://www.linkedin.com/in/magnus-...ofile:ACoAAAgW3tQBaAlEnmRiNA0C_4jsXoq4ncLf-Zg
Magnus Östberg
Magnus ÖstbergSoftware Officer at Mercedes-Benz AG
1 Tag

Three to four hours a week – that’s the average amount of time I invest in podcasts 🎧 And our #MercedesBenz Podcast #FutureDimensions is one of them. If you’ve always wondered what the future looks (& sounds) like, this podcast is for you.

The current episode called “What if AI let us grow beyond ourselves?” has especially caught my attention, because #ChatGPT is one of the interview partners. How does that work? Listen for yourself: https://lnkd.in/exuCM5eE
 
  • Like
  • Love
  • Fire
Reactions: 16 users

wilzy123

Founding Member
I wasn't buying a current connection between MB and Brainchip on the basis MB hadn't deleted an article on their website about the EQXX.
.....Including all of the facts and speculation served up to you on a platter here either by the sounds of it. LMAO.🤡🤡🤡🤡🤡
 
  • Like
  • Haha
  • Love
Reactions: 12 users

TECH

Regular
With Akida 2 due to make it's presence late in this quarter (not a given as such), don't forget about the first child, AKD 1, despite this brilliant achievement now being 4 years old, there are still exciting possibilities that I believe the company is still exploring with that IP and chip.

To say that we aren't heading in the right direction would be a total lie, keeping in mind, the clients we are working with obviously have
the final say as to whether they proceed with a design or not, controlling what goes on within Brainchip's offices worldwide is clearly what
the focus is on, and it equates to hard work, it is a mentally draining type of work that is driven by passion and the desire to create and ultimately
succeed at the highest level.

Unless you are a trader (daily) or short of some capital to pay some bills for example, why would you even consider off-loading your
precious shares, it's a personal choice, I get it, just imagine how patient all the staff are whom are driving Brainchip forward day in day out,
without all that brainpower we wouldn't even have a heartbeat...I agree with a number of posters, HVAC is going to be Renesas' product
launch into the Edge AI market, it matches up with that interview I posted last week, all will be revealed this year hopefully.

Have a good Sunday.....Tech :coffee:(y)
 
  • Like
  • Love
  • Fire
Reactions: 47 users

Iseki

Regular
The trouble is that Renesas have only licensed 2 nodes.
But all the fancy new stuff that Akida2 can bring (Transformers etc) requires 8 nodes.

So who will build an Akida-p?
What version of Akida is Renesas building?

You can't ask BRN, because they only say it is a Reneas chip.

Sad.
 
  • Like
  • Sad
Reactions: 5 users

Slade

Top 20
The trouble is that Renesas have only licensed 2 nodes.
But all the fancy new stuff that Akida2 can bring (Transformers etc) requires 8 nodes.

So who will build an Akida-p?
What version of Akida is Renesas building?

You can't ask BRN, because they only say it is a Reneas chip.

Sad.
Keep trying. But please don’t use too much oxygen.
 
  • Haha
  • Like
  • Love
Reactions: 25 users

Cartagena

Regular
Hi, greeting to everyone. New contributor here and also long term holder of this company. The Edge Computing Expo will be taking place in Amsterdam in September this year. Would anyone know if Brainchip is attending? I looked on the site and didn't find the brainchip logo possibly we could be however this is the "place to be" amongst the other multi nationals .
 

Attachments

  • Screenshot_20230716-123234.png
    Screenshot_20230716-123234.png
    1.9 MB · Views: 47
  • Like
Reactions: 2 users

TheFunkMachine

seeds have the potential to become trees.
https://www.linkedin.com/posts/ai-l...4-o_7m?utm_source=share&utm_medium=member_ios

Brainchip and AI labs partnership article says this:

A combination of the Minsky AI Engine, smart models, and the sensory inference capabilities of BrainChip’s Akida™ provides a compelling and cost-effective solution to these real-world problems.

“Minsky is the world’s most easy-to-use, proprietary, enterprise-grade AI engine that does not require data scientists or software developers to model and run predictions,” said Bhasker Rao, Founder of Ai Labs. “When optimized for BrainChip’s neuromorphic IP platform, the resulting solution will fuel rapid growth in the industrial AIoT space. Together with BrainChip’s Akida IP and MetaTFTM, we can radically improve development and deployment of predictive maintenance applications, such as vibration analysis, gas sensing, temperature sensing and others, preventing costly potential failures and reducing machine downtime. This streamlines processes and increases industrial efficiency and profitability.”

Additionally, since similar machines do not all fail the same way or at the same rate, models may have to be customized for different production lines. This is where Akida’s on-device learning capability provides a differentiated and secure approach to intelligent customization.

“Technological advancements in industrial infrastructure such as oil exploration, manufacturing, mining, transportation to mention a few, have enabled some amazing features that make for a better customer experience, but they also have created additional points of failure that can be difficult to assess,” said Rob Telson, VP of Ecosystems and Partnerships at BrainChip. “Using a combination of AI Labs, the Minsky AI Engine and our Akida platform, we are providing a way to intelligently identify points of failure, predict issues and proactively fix them before they happen. This intelligence is a game changer, and we are proud to partner with AI Labs to bring this revolutionary technology to market.”

The infamous Rob like strikes again:)
 

Attachments

  • 6A5774D9-181D-4C7A-99C7-409BF75510C7.jpeg
    6A5774D9-181D-4C7A-99C7-409BF75510C7.jpeg
    1.4 MB · Views: 121
  • Like
  • Fire
  • Love
Reactions: 45 users

TheFunkMachine

seeds have the potential to become trees.
Just back to Valeo.

Posted this about this time last year with some of the authors being Valeo.

Were playing with CNN2SNN simulations.

Full paper in original post & I added the conclusion snip this time.



Don't know if been posted already, but skimming some posts around Valeo earlier and this paper released recently, June 6th.

Paper attached.

At least we have Valeos name for it...Spiking Lidar......SpikiLi :D

Like to think they just replicating / simulating what Akida can already do for their own confirmations as part of the process.

No direct reference to us though :cautious:

Maybe @Diogenese will understand the implications better and explain in simpler terms for us.

IMG_20220614_215425.jpg



IMG_20220614_220217.jpg


View attachment 40027
I just did some Linked in searches on the names on this article you posted and thought I would share them with you. One of them has a strong connection to Qualcomm LiDAR which I found interesting and another is from CEA Leti with focus on SNN.

I’m not aware of any other company that can preform CNN to SNN covertion which makes me hopeful that brainchip is a key in all of this but it’s hard to say.
 

Attachments

  • 24713857-C31F-4134-835C-A93BF2694B0A.png
    24713857-C31F-4134-835C-A93BF2694B0A.png
    815 KB · Views: 91
  • 5C746CAB-8038-473D-8598-7C0D685D0D90.png
    5C746CAB-8038-473D-8598-7C0D685D0D90.png
    469.4 KB · Views: 79
  • F32D4B68-3FC4-4475-98FA-B6EBC6E7257C.png
    F32D4B68-3FC4-4475-98FA-B6EBC6E7257C.png
    388.8 KB · Views: 83
  • 43140E32-863B-46AD-8088-921DCB34A939.png
    43140E32-863B-46AD-8088-921DCB34A939.png
    622.8 KB · Views: 73
  • Like
  • Fire
  • Love
Reactions: 29 users
https://www.linkedin.com/posts/ai-l...4-o_7m?utm_source=share&utm_medium=member_ios

Brainchip and AI labs partnership article says this:

A combination of the Minsky AI Engine, smart models, and the sensory inference capabilities of BrainChip’s Akida™ provides a compelling and cost-effective solution to these real-world problems.

“Minsky is the world’s most easy-to-use, proprietary, enterprise-grade AI engine that does not require data scientists or software developers to model and run predictions,” said Bhasker Rao, Founder of Ai Labs. “When optimized for BrainChip’s neuromorphic IP platform, the resulting solution will fuel rapid growth in the industrial AIoT space. Together with BrainChip’s Akida IP and MetaTFTM, we can radically improve development and deployment of predictive maintenance applications, such as vibration analysis, gas sensing, temperature sensing and others, preventing costly potential failures and reducing machine downtime. This streamlines processes and increases industrial efficiency and profitability.”

Additionally, since similar machines do not all fail the same way or at the same rate, models may have to be customized for different production lines. This is where Akida’s on-device learning capability provides a differentiated and secure approach to intelligent customization.

“Technological advancements in industrial infrastructure such as oil exploration, manufacturing, mining, transportation to mention a few, have enabled some amazing features that make for a better customer experience, but they also have created additional points of failure that can be difficult to assess,” said Rob Telson, VP of Ecosystems and Partnerships at BrainChip. “Using a combination of AI Labs, the Minsky AI Engine and our Akida platform, we are providing a way to intelligently identify points of failure, predict issues and proactively fix them before they happen. This intelligence is a game changer, and we are proud to partner with AI Labs to bring this revolutionary technology to market.”

The infamous Rob like strikes again:)
AI Labs have 158 followers, post on Linkedin about once a day, and the only likes they receive are from our clan of devout BRN followers.
It would be good if someone (who is more tech savvy than me), perhaps get in touch with AI Labs, and see if they can reference Akida in their posts, that is if they are using it in the product they are promoting. That would surely get their name put out, as well as ours, linked to a product..

Can be a win-win for all, but then again the secretive NDA topic could get in the way.
 
  • Like
  • Fire
Reactions: 9 users

Kachoo

Regular
I just did some Linked in searches on the names on this article you posted and thought I would share them with you. One of them has a strong connection to Qualcomm LiDAR which I found interesting and another is from CEA Leti with focus on SNN.

I’m not aware of any other company that can preform CNN to SNN covertion which makes me hopeful that brainchip is a key in all of this but it’s hard to say.
Not sure about every connection.

But my key take away is many many have been researching and studying SNN with possible links.

One of the challenges BRN has had as the pioneer was convincing clients to change well as more and more see benifits and keep forging SNN BRN will have more ease in selling to more willing to adapt even with competition. The value of SNN investments clearly is starting to show.

There is yet an article that I have seen to prove or dispute Nueromorphic computing is not going to impact the world mist articles boost about benifits. Accuracy was one of the issues I read about but BRN has overcome that issue for the most part.
 
  • Like
  • Fire
Reactions: 19 users

Cartagena

Regular
https://www.linkedin.com/posts/ai-l...4-o_7m?utm_source=share&utm_medium=member_ios

Brainchip and AI labs partnership article says this:

A combination of the Minsky AI Engine, smart models, and the sensory inference capabilities of BrainChip’s Akida™ provides a compelling and cost-effective solution to these real-world problems.

“Minsky is the world’s most easy-to-use, proprietary, enterprise-grade AI engine that does not require data scientists or software developers to model and run predictions,” said Bhasker Rao, Founder of Ai Labs. “When optimized for BrainChip’s neuromorphic IP platform, the resulting solution will fuel rapid growth in the industrial AIoT space. Together with BrainChip’s Akida IP and MetaTFTM, we can radically improve development and deployment of predictive maintenance applications, such as vibration analysis, gas sensing, temperature sensing and others, preventing costly potential failures and reducing machine downtime. This streamlines processes and increases industrial efficiency and profitability.”

Additionally, since similar machines do not all fail the same way or at the same rate, models may have to be customized for different production lines. This is where Akida’s on-device learning capability provides a differentiated and secure approach to intelligent customization.

“Technological advancements in industrial infrastructure such as oil exploration, manufacturing, mining, transportation to mention a few, have enabled some amazing features that make for a better customer experience, but they also have created additional points of failure that can be difficult to assess,” said Rob Telson, VP of Ecosystems and Partnerships at BrainChip. “Using a combination of AI Labs, the Minsky AI Engine and our Akida platform, we are providing a way to intelligently identify points of failure, predict issues and proactively fix them before they happen. This intelligence is a game changer, and we are proud to partner with AI Labs to bring this revolutionary technology to market.”

The infamous Rob like strikes again:)
Good article TFM. I like to think that's a concrete collaboration .. Any partnerships between SynSense (SS) and Brainchip? I am aware our partner Prophesee and SS are partners with eachother. SS are launching their new chip Speck, also an edge AI chip (our close competitor). It is being commercialised as per their website and media they released. Would like to find out more hopefully about any connection we have with SS? AI is moving fast
 
  • Like
Reactions: 2 users

Cartagena

Regular
Hi, greeting to everyone. New contributor here and also long term holder of this company. The Edge Computing Expo will be taking place in Amsterdam in September this year. Would anyone know if Brainchip is attending? I looked on the site and didn't find the brainchip logo possibly we could be however this is the "place to be" amongst the other multi nationals .
Can anyone see my posts??
 
  • Like
Reactions: 2 users
Top Bottom