BRN Discussion Ongoing

[NjeDK]

Unlocking New Real-world Generative AI Use Cases on the Mobile CPU

Armv9 CPU technologies are unleashing innovative generative AI use cases that run entirely on today’s AI-enabled flagship smartphone devices.

newsroom.arm.com

Have anyone seen this from ARMs website?

Could it be Akida in their CPU?

 

[Frangipani]

Interesting like of a BrainChip post on LinkedIn by a Fraunhofer ITWM (Institut für Techno*- und Wirtschaftsmathematik / Institute for Industrial Mathematics) research scientist, given this Fraunhofer Institute’s High Performance Computing division is co-coordinating a project called STANCE (Strategic Alliance For Neuromorphic Computing and Engineering) alongside Fraunhofer IIS (Institut für Integrierte Schaltungen / Institute for Integrated Circuits), which aims to push for the adoption of spiking and neuromorphic technologies in industrial production by bringing together users and solution providers. The STANCE project got underway five months ago.

*Ich musste echt zweimal hinschauen, aber die forschen wohl doch nicht über Raves…
@cosors, ohne Bindestrich wäre das ansonsten doch echt der perfekte Arbeitsplatz für dich gewesen, oder?!

 

[Frangipani]

Please come by the AI Hardware & Edge AI Summit and check out the latest advancements with BrainChip! | Edward Lien, Executive MBA

Please come by the AI Hardware & Edge AI Summit and check out the latest advancements with BrainChip! We'll be at booth #58. See you there.

www.linkedin.com

View attachment 68967 View attachment 68968 View attachment 68969 View attachment 68970

Speaking of LinkedIn likes:

Fair chance this catch-up between our new Regional Sales Manager in Taiwan and the gentleman working for NVIDIA is going to be more of a private nature than talking business - they both worked for AMD around the same time and Jackie Soussan used to organise overseas sales meetings, including one or several in Taiwan. I wouldn’t read too much into it.


Edward Lien:

Jackie Soussan:

 

[IloveLamp]

Speaking of LinkedIn likes:

Fair chance this catch-up between our new Regional Sales Manager in Taiwan and the gentleman working for NVIDIA is going to be more of a private nature than talking business - they both worked for AMD around the same time and Jackie Soussan used to organise overseas sales meetings, including one or several in Taiwan. I wouldn’t read too much into it.


Edward Lien:


View attachment 69153


Jackie Soussan:

View attachment 69154

 

[Frangipani]

View attachment 69155

Wrong. I am not a guy, to start with.

You, however, seem to be more of a “glass is overflowing” kind of guy? Right?

 

[IloveLamp]

Wrong. I am not a guy, to start with.

You, however, seem to be more of a “glass is overflowing” kind of guy? Right?

Over the past couple of years I've found more than 30-40 likes (at least) both ways on LinkedIn between nvidia and brn so yes, i believe there's more to it

Dyor

The fact you are a woman makes SO MUCH SENSE right now, thanks for sharing

 

[7für7]

ASX announcement as usual

 

[IloveLamp]

Just reposting that thumbnail image of yours in full size, as it reveals the name of a new demo on the right monitor as well as on the info leaflet next to it:


View attachment 69108


View attachment 69109


It says: “LLMs on the Edge with RAG” resp. “LLM with RAG on Akida”

I googled RAG, which means “Retrieval-Augmented Generation”.

Retrieval-augmented generation - Wikipedia

en.m.wikipedia.org

“Retrieval augmented generation (RAG) is a type of generative artificial intelligence that has information retrieval capabilities. It modifies interactions with a large language model (LLM) so that the model responds to user queries with reference to a specified set of documents, using this information in preference to information drawn from its own vast, static training data. This allows LLMs to use domain-specific and/or updated information.[1] Use cases include providing chatbot access to internal company data, or giving factual information only from an authoritative source.[2]”

Learn more about Multimodal RAG from our very own Vasudev Lal in the latest DeepLearning.AI course. | Intel Labs

Learn more about Multimodal RAG from our very own Vasudev Lal in the latest DeepLearning.AI course. Andrew Ng explains more about what to expect in the video below. #DeepLearning #LVLM #RAG

www.linkedin.com

 

[itsol4605]

Nice !! Is there any reason for it ??

 

[TheDon]

Nice !! Is there any reason for it ??

View attachment 69159

Wow!it's pointing towards 5 trillion market Cap. I'm excited!

TheDon

 

[7für7]

Nice !! Is there any reason for it ??

View attachment 69159

thank me later buddy … I know it’s hard for beginner to understand the system… here is a simple guide… don’t overdo it! Step by step

 

[Diogenese]

View attachment 69146

Hi @Frangipani ,

The excerpt I posted immediately below the post from which you inferred a suggested Akida link refers to an algorithm.

Further on, that article continues:

In artificial intelligence, machine learning is a way that computers learn from data—lots of data—to identify patterns and make decisions or draw conclusions.

This automated process can be powerful when the patterns might not be obvious to human researchers looking at the same data, which is typical for large, complex data sets such as those involved in imaging and spectral analysis.

In MOMA's case, researchers have been collecting laboratory data for more than a decade, according to Victoria Da Poian, a data scientist at NASA Goddard who co-leads development of the machine learning algorithm. The scientists train the algorithm by feeding it examples of substances that may be found on Mars and labeling what they are. The algorithm will then use the MOMA data as input and output predictions of the chemical composition of the studied sample, based on its training.

"The more we do to optimize the data analysis, the more information and time scientists will have to interpret the data," Da Poian said. "This way, we can react quickly to results and plan next steps as if we are there with the rover, much faster than we previously would have."

The collected data will be used in building various AI models which can be used by the AI algorithm for classification or "labelling" as well as for navigation. If NASA are using a conventional CNN algorithm, this will be a significant drain on their elctricity supply.

BRN has been engaged with NASA for some years. It would be surprising if Victoria Da Poian were to be unaware of Akida. In particular NASA would have been aware of TeNNs for a couple of years.

The rover must navigate autonomously, and Akida has famously been found to enable an increase in cruising speed of such rovers due to its real time processing capabilities. The rover is said to have a range of several km.

While Akida has been radhardened by Vorago, there is nothing to suggest the Akida SoC is incorporated in the rover. Indeed, as you point out, time would be against this possibility ... but next time?

Sean Hehir recently acknowledged that BRN now has an additional product line for algorithms.

Rosalind is not due to be launched til 2028, so there is time to incorporate TeNNs/Akida 2 algorithms in the Mars rover. But who knows? It may already be there.

FN:
https://www.esa.int/Science_Explora...th_of_ESA_s_ExoMars_Rosalind_Franklin_mission

FAQ: The ‘rebirth’ of ESA’s ExoMars Rosalind Franklin mission​

13/03/2023
...
There is a plan to re-use in the new Lander the qualified onboard computer, the radar doppler altimeter and the parachute system that had been developed for the previous version of the mission.
...

This refers to the lander module, but the rover will also have a "qualified" processor which will not be readily replaced. However:

Any obsolescence of parts will be taken care of with appropriate refurbishments during the development of the ExoMars Rosalind Franklin Mission.

 

[Gazzafish]

Posted just yesterday. Very interesting if you zoom in

x.com

x.com

 

[davidfitz]

Nice !! Is there any reason for it ??

View attachment 69159

Easy to get excited with any positive move but unfortunately, it's just the same games being played day in day out.

 

[Labsy]

Nice !! Is there any reason for it ??

View attachment 69159

Battle of the bots... The powers that be want an increase now for who knows why.... Maybe coz they successfully bought a bunch at 15.5 cents again...and want to sell at 30 cents again

 

[Pom down under]

Hopefully not to long now to we get a real announcement, but must admit I’m a bit fed up with the wait.

 

[Mt09]

Is the evaluation kit new on megachips site?

Edge AI | MegaChips Corporation

We respond to various issues of customers including cost and development period with our integrated support leveraging foundries around the world.

www.megachips.co.jp

Been on there a fair while from memory.

 

[Diogenese]

ESA started this project in January 2023:

Continual online learning with spiking neural networks​

https://www.esa.int/gsp/ACT/projects/continual-online-learning/

...
In contrast, current AI systems are almost exclusively trained on a particular task using a large amount of data collected prior to training, so everything about the data (type, statistical distribution and so on) is known prior to the training phase. Therefore, the ability to continually adapt and learn is perhaps one of the most profound points of distinction between current AI systems and biological organisms.

An interesting thing about Akida 2 is that the Akida NPU NN does not use backpropagation (BP), (BP) is used by TeNNs.

It's easy to see why ESA are so enthused by Akida.

 

[Diogenese]

Nice tablecloth…

I prefer the tailored look to the puckered end drape.

 

[davidfitz]

I wonder what will be discussed at this?

SPAICE – SPAICE Conference

spaice.esa.int