BRN Discussion Ongoing

greatlake

Regular
Not sure if This was posted already.. but anyway interesting. It was posted from a user in the German forum

Thanks for posting this. Every major sensor player should be highly interested in this. Akida being sensor agnostic coupled with generation 2.0 will open many more doors.

After all, that’s what we do. Make sensors smart.
 
  • Like
  • Fire
  • Love
Reactions: 34 users

Cgc516

Regular
Has anyone heard ADLink Taiwanese company?
Edge computing—doesn’t say SNN, but also partnered with Intel , Nvidia, Arm, Qualcomm .
 
  • Like
  • Thinking
  • Fire
Reactions: 5 users

7für7

Top 20
Smart people in Singapur right?

 
  • Like
  • Fire
  • Wow
Reactions: 19 users

Frangipani

Regular
Looks to me like the different charges went bang and we've lost both?
Hopefully I'm wrong.
Frangi, I would also miss your detective work for us very much.

Don’t worry, my detective agency has not closed down. 🕵️‍♀️ So stay tuned!

But I do have a life outside TSE, and yesterday, I preferred to enjoy real-life fresh spring air over breathing in all those toxic fumes blowing in my direction in the virtual world of anonymous posters.
 
  • Like
  • Love
  • Fire
Reactions: 22 users

RobjHunt

Regular
Simon Thorpe has advocated a universal income for many years to offset the disruption to the workplace from AI.

Now Sam Altman has jumped on the bandwagon and will no doubt make a bigger splash:


OpenAI's Statement SHOCK the Entire Industry! AI Riots vs "Moore's Law for Everything" by Sam Altman (youtube.com)



OpenAI's Statement SHOCK the Entire Industry! AI Riots vs "Moore's Law for Everything" by Sam Altman

...
@ 19:30

View attachment 58361


This will be a hard sell in ideologically entrenched America. Still they hounded Chaplain out so this bloke shouldn't prove too difficult.

Ole mate Dodgey, your posts are nothing but astounding.
 
  • Like
  • Love
Reactions: 11 users

RobjHunt

Regular
  • Like
  • Haha
Reactions: 2 users
Saw a couple of our employees "liking" mamba and wondered what it was.

Couple snips below and curious whether it was a general "like", great development thing or whether it is something they are now working on too.

Is possible the PeaBrane/Mamba-Tiny on Git and liked is something that Rudy created by the looks from the full Mamba but maybe I'm reading it wrong?

Interesting none the less I think given the Ai in 24 links Mamba, Transformers & Neuromorphic all together.



Mamba: This refers to the emergence of new, ground breaking AI architectures like transformers and neuromorphic computing. These architectures mimic the human brain’s structure and function, allowing for significantly faster processing and deeper learning capabilities. Mamba-based models will revolutionize areas like natural language processing, image recognition, and robotics.

  • Mamba Architecture: Revolutionizing Sequence Modelling
    1. Mamba, a ground breaking architecture, represents a leap forward from the Transformer models.
    2. It addresses the computational challenges of large-scale sequence processing.
    3. Albert Goo’s work on structured state spaces inspired Mamba’s development.
    4. The architecture’s potential lies in its ability to handle extensive sequences, as demonstrated in DNA classification tasks.


[Submitted on 1 Dec 2023]

Mamba: Linear-Time Sequence Modelling with Selective State Spaces​

Albert Gu, Tri Dao
Foundation models, now powering most of the exciting applications in deep learning, are almost universally based on the Transformer architecture and its core attention module. Many subquadratic-time architectures such as linear attention, gated convolution and recurrent models, and structured state space models (SSMs) have been developed to address Transformers' computational inefficiency on long sequences, but they have not performed as well as attention on important modalities such as language. We identify that a key weakness of such models is their inability to perform content-based reasoning, and make several improvements. First, simply letting the SSM parameters be functions of the input addresses their weakness with discrete modalities, allowing the model to selectively propagate or forget information along the sequence length dimension depending on the current token. Second, even though this change prevents the use of efficient convolutions, we design a hardware-aware parallel algorithm in recurrent mode. We integrate these selective SSMs into a simplified end-to-end neural network architecture without attention or even MLP blocks (Mamba). Mamba enjoys fast inference (5× higher throughput than Transformers) and linear scaling in sequence length, and its performance improves on real data up to million-length sequences. As a general sequence model backbone, Mamba achieves state-of-the-art performance across several modalities such as language, audio, and genomics. On language modeling, our Mamba-3B model outperforms Transformers of the same size and matches Transformers twice its size, both in pretraining and downstream evaluation.




1709470139376.png




1709471145144.png






Rudy Pei

Physicist | ML researcher | quantum & neuromorphic computing | behavioral economics | composer
3w Edited

Mamba is a new state-space model out-performing transformers on "everywhere tried". Originally, it was trained with associative scan, which pytorch does not support natively, hence the need for custom CUDA kernels. However, there is a simple math trick to express the associative scans used in mamba as a ratio of two cumulative sums. This makes for an efficient native pytorch implementation of mamba possible. How? Check out my simple repo with an one-file implementation of this idea forking from the mamba-minimal repo https://lnkd.in/g5QR7yHC #mamba #llm
GitHub - PeaBrane/mamba-tiny: Simple, minimal implementation of the Mamba SSM in one file of PyTorch. More efficient than the minimalist version but less efficient than the original mamba implementation.

GitHub - PeaBrane/mamba-tiny: Simple, minimal implementation of the Mamba SSM in one file of PyTorch. More efficient than the minimalist version but less efficient than the original mamba implementation.

github.com

 
  • Like
  • Thinking
  • Fire
Reactions: 23 users

Sirod69

bavarian girl ;-)
I tried to translate for you, it's in Italian. Since it is our partner I wanted to post it here.😘

MYWAI, the cutting-edge start-up in the field of #EdgeAI and #GenAI, will be present at the round table of the Smart Manufacturing Summit, organized by The Innovation Group, in #Milan on 9-10 April 2024.

🔍What does the use of AI imply for the Industry 5.0 Transition?

- #EdgeAI: AI algorithms operating directly on #edge computers or in intelligent neuromorphic sensors.
- #GenAI: able to generate the data missing from your dataset to train maintenance and quality control algorithms
-#ReinforcementLearning and #PhysicsInformedNetworks for the intelligent control of multiphysical processes and production machinery
-Semantic and 3d #DigitalTwins based on open standards such as the new #OPCUA Companion Specs and #NeuralRendering 3d
-#MultimodalAI to fuse audio, video and image data with plant and machinery time series for next-generation simulators and decision-making supports...

 
  • Like
  • Love
  • Fire
Reactions: 32 users
Smart people in Singapur right?


There's no way to stop, the ever swelling ranks of the "Useless Class".
A populous, that is not only unemployed, but unemployable.

Already the number of people in Australia, on support is high.
"In total, an estimated 38.7% of the population aged 15–64 was receiving government income support in mid-2020"

Another close to 10% are direct government employees.
"There were 2,430,400 public sector employees in the month of June 2023 comprising: 350,300 employees in Commonwealth government (including defence force personnel); 1,871,900 in state government; and. 208,200 in local government".

Not to mention pensions..
"As at 31 March 2023, around 2.8 million people aged 65 and over received income support payments, equating to 63% of the population aged 65 and over. Of these, the vast majority (92% or 2.6 million) received Age Pension".

Most developed countries, would be similar, is it any wonder why government debt is spiraling out of control, practically everywhere?


No amount of education, will be able to compete with the type of A.I. they are talking about.

Only highly specialised jobs, trades and services (which include everything from hairdressing, to cleaning etc) will remain.

The vast majority of the business cohort, office jobs etc. Will be ripe for obsolescence, within the next 5 to 10 years (at the current rate of advancement IMO).

There's often the argument, that people will just learn new skills etc..

That is BS, in my humble opinion.

Everyone, who invests in the changes that are happening, will come out much better than most, when all is said and done.

You're either on the ship to progress, or you're in the water clinging to a government life preserver, with just enough money for food and online entertainment (opiate of the masses).
Which many will and are, already satisfied with.

That's the Future, the youth of today, have to look forward to.

For some, it will be their Dream Lifestyle.


20240304_010546.jpg
 
Last edited:
  • Like
  • Fire
  • Wow
Reactions: 18 users
Whilst only a small mention, great to see confirmation we're still in the mix with some heavy hitters over at NASA & Ames Research Centre.

Gotta say something about the tech :)

The original neuromorphic flight "test" was with Loihi and the TechEd Sat from what I read but states moving to next phases with more neuromorphic.

BRAINSTACK – A Platform for Artificial Intelligence & Machine Learning Collaborative Experiments on a Nano-Satellite

Date Acquired
August 2, 2023
Publication Date
August 11, 2023
Publication Information
Publication: SmallSat Conference Proceedings

View attachment 42347

View attachment 42348

Part of the conclusion...

Building on earlier successes with GPUs, and more recently, a neuromorphic processor flight test, the notion of a collaborative BrainStack orbital AI/ML
laboratory module is presented. The intention is to be able to perform experiments on multiple hardware and
software AI/ML elements on the same flight with different collaborative teams.

The TES-n Common AI/ML Software Interface will permit a menu driven set
of experiments across individual elements. The next set of TES BrainStack experiments will host combinations
of GPUs and neuromorphic processors, with flexibility
to support upcoming novel systems and their unique interfaces. Such a collaborative BrainStack system will greatly expand the use of these remarkable new tools and methods in the space sector.


Full paper:

HERE
Whilst I originally posted the NASA Brainstack paper back in Aug, I just found the presso from the same time while doing some random googling.

What I found interesting revisiting the table, which I hadn't thought of before, is which ones have been tested and appears to have a loose order via the Sat programs / launches.

If they stay true to that format then could we expect we could be considered next in line :unsure:

Presso HERE


IMG_20240303_222339.jpg
 
  • Like
  • Love
  • Fire
Reactions: 36 users
Probs posted but don't recall seeing it or I missed it, but a quick chat with RT at CES.

Not much given away and interviewer was a little meh for mine.

Did like the little run through bit by RT of what Merc are doing with in cabin...plug for a friend or something more :unsure:

RT tried hard to get some stuff out there, but guess time becomes a consideration so seemed a little rushed.


 
  • Like
  • Fire
  • Love
Reactions: 17 users
  • Like
  • Love
  • Fire
Reactions: 31 users

IloveLamp

Top 20
  • Like
  • Thinking
  • Love
Reactions: 13 users

Iseki

Regular
Don’t worry, my detective agency has not closed down. 🕵️‍♀️ So stay tuned!

But I do have a life outside TSE, and yesterday, I preferred to enjoy real-life fresh spring air over breathing in all those toxic fumes blowing in my direction in the virtual world of anonymous posters.
You're a bad smell frangipani. Stay away until those you targeted feel free to return.
 
  • Like
  • Love
  • Sad
Reactions: 14 users

7für7

Top 20
Good Morning!
Can I see a BOOM please? Thank you!
 
  • Like
  • Fire
  • Love
Reactions: 13 users

skutza

Regular
You're a bad smell frangipani. Stay away until those you targeted feel free to return.
Maybe this type of message is best sent through messages so we don't all have to read it? Or do you need team support for this type of comment? (Asking for a friend).
 
  • Like
  • Love
  • Fire
Reactions: 17 users

Tony Coles

Regular
  • Like
  • Love
  • Fire
Reactions: 33 users

Iseki

Regular
Maybe this type of message is best sent through messages so we don't all have to read it? Or do you need team support for this type of comment? (Asking for a friend).
Maybe tell your friend to follow their own advice. Frankly, if they can't be more generous and less of a narcissist, if they can't be slightly distracted by the improving SP, then move on. Why do they want to target individual contributors anyway?
 
  • Like
  • Thinking
  • Wow
Reactions: 5 users

Labsy

Regular
Goodmorning everyone.... Have a great week! ;) It's going to be a cracker.... ( just hoping of course, no insider info). 🚀🚀🚀🙏🙏🙏🔥🔥🔥🍻
 
  • Like
  • Fire
  • Haha
Reactions: 26 users

mrgds

Regular
Goodmorning everyone.... Have a great week! ;) It's going to be a cracker.... ( just hoping of course, no insider info). 🚀🚀🚀🙏🙏🙏🔥🔥🔥🍻
A "supercut " of Intels presentation with "ARM" partnership worth a watch imo.

Now what little start-up is partnered with both ? ............:unsure: ..........................................................;)



AKIDA BALLISTA
 
  • Like
  • Fire
  • Love
Reactions: 37 users
Top Bottom