BRN Discussion Ongoing

[AARONASX]

hmmm maybe you missed the Linked post from their official account

“come on guys.. it was an exhibition where companies could check out our amazing technology.. we took some pictures so what? It’s not our fault you interpret everything aa “akida inside”. != "We welcomed partners like PROPHESEE, Edge Impluse, and RTX..."


View attachment 76123

I cannot say for sure just pure speculation like everything else.... Tony's statement was NO partnership between RTX and Brainchip.....that does not mean they are not a customer via a different channel, eg Megachips, etc. maybe still a end user. IMO

 

[Bravo]

Howdy All,

So I was just having bit of a poke around to see if I could find out a bit more about the Klepsydra and Frontgrade Gaisler collaboration as per the above post and I stumbled on a Klepsydra Linkedin post from a couple of months ago which mentioned the "REBECCA Project" (see below).

I did a Google search and found the website for the REBECCA Project, which sure enough mentions neuromorphic computing.

As you can see from the screen shots from the website, the REBECCA project is co-funded by the European Union. The aim of the project is to democratize the development of edge AI systems including a hardware and software stack centred around a RISC-V CPU.

As described on the website, it will develop a novel chip that will include a neuromorphic AI accelerator, which you would think would be very likely to be AKIDA, since Frontgrade Gaisler has taken a commercial license for the Akida neural processor IP.

I also looked for further information on the Grant agreement ID: 101097224 and have posted a screen shot of that as well. Total cost for the project is €8 498 328,59 (with EU contribution comprising €2 744 319,72)!!!! The end date is set at 31 July 2026.

Clearly this is a very big deal because it says on the Cordis summary that the "project will contribute to realising business and societal opportunities by validating and demonstrating its approach on real-world use cases and benchmarks based on real-world applications from the smart appliances, energy generation, infrastructure inspection, avionics, automotive and health domains".


View attachment 76117


View attachment 76110


View attachment 76119




View attachment 76115

View attachment 76116

About Rebecca | Rebecca

www.rebecca-chip.eu

View attachment 76120

About Rebecca | Rebecca

www.rebecca-chip.eu

Reconfigurable Heterogeneous Highly Parallel Processing Platform for safe and secure AI | REBECCA Project | Fact Sheet | HORIZON | CORDIS | European Commission

REBECCA, a heavily SME-driven project, will democratize the development of novel edge AI systems. Towards this aim, REBECCA will develop a purely European complete Hardware(HW) and Software(SW) stack around a RISC-V CPU, which will provide significantly higher levels of a)...

cordis.europa.eu

To aid in understanding the potential connections proposed above....

 

[MDhere]

Howdy All,

So I was just having bit of a poke around to see if I could find out a bit more about the Klepsydra and Frontgrade Gaisler collaboration as per the above post and I stumbled on a Klepsydra Linkedin post from a couple of months ago which mentioned the "REBECCA Project" (see below).

I did a Google search and found the website for the REBECCA Project, which sure enough mentions neuromorphic computing.

As you can see from the screen shots from the website, the REBECCA project is co-funded by the European Union. The aim of the project is to democratize the development of edge AI systems including a hardware and software stack centred around a RISC-V CPU.

As described on the website, it will develop a novel chip that will include a neuromorphic AI accelerator, which you would think would be very likely to be AKIDA, since Frontgrade Gaisler has taken a commercial license for the Akida neural processor IP.

I also looked for further information on the Grant agreement ID: 101097224 and have posted a screen shot of that as well. Total cost for the project is €8 498 328,59 (with EU contribution comprising €2 744 319,72)!!!! The end date is set at 31 July 2026.

Clearly this is a very big deal because it says on the Cordis summary that the "project will contribute to realising business and societal opportunities by validating and demonstrating its approach on real-world use cases and benchmarks based on real-world applications from the smart appliances, energy generation, infrastructure inspection, avionics, automotive and health domains".


View attachment 76117


View attachment 76110


View attachment 76119




View attachment 76115

View attachment 76116

About Rebecca | Rebecca

www.rebecca-chip.eu

View attachment 76120

About Rebecca | Rebecca

www.rebecca-chip.eu

Reconfigurable Heterogeneous Highly Parallel Processing Platform for safe and secure AI | REBECCA Project | Fact Sheet | HORIZON | CORDIS | European Commission

REBECCA, a heavily SME-driven project, will democratize the development of novel edge AI systems. Towards this aim, REBECCA will develop a purely European complete Hardware(HW) and Software(SW) stack around a RISC-V CPU, which will provide significantly higher levels of a)...

cordis.europa.eu

Nothing short of amazing bravo! And that's a wrap.

 

[Diogenese]

Howdy All,

So I was just having bit of a poke around to see if I could find out a bit more about the Klepsydra and Frontgrade Gaisler collaboration as per the above post and I stumbled on a Klepsydra Linkedin post from a couple of months ago which mentioned the "REBECCA Project" (see below).

I did a Google search and found the website for the REBECCA Project, which sure enough mentions neuromorphic computing.

As you can see from the screen shots from the website, the REBECCA project is co-funded by the European Union. The aim of the project is to democratize the development of edge AI systems including a hardware and software stack centred around a RISC-V CPU.

As described on the website, it will develop a novel chip that will include a neuromorphic AI accelerator, which you would think would be very likely to be AKIDA, since Frontgrade Gaisler has taken a commercial license for the Akida neural processor IP.

I also looked for further information on the Grant agreement ID: 101097224 and have posted a screen shot of that as well. Total cost for the project is €8 498 328,59 (with EU contribution comprising €2 744 319,72)!!!! The end date is set at 31 July 2026.

Clearly this is a very big deal because it says on the Cordis summary that the "project will contribute to realising business and societal opportunities by validating and demonstrating its approach on real-world use cases and benchmarks based on real-world applications from the smart appliances, energy generation, infrastructure inspection, avionics, automotive and health domains".


View attachment 76117


View attachment 76110


View attachment 76119




View attachment 76115

View attachment 76116

About Rebecca | Rebecca

www.rebecca-chip.eu

View attachment 76120

About Rebecca | Rebecca

www.rebecca-chip.eu

Reconfigurable Heterogeneous Highly Parallel Processing Platform for safe and secure AI | REBECCA Project | Fact Sheet | HORIZON | CORDIS | European Commission

REBECCA, a heavily SME-driven project, will democratize the development of novel edge AI systems. Towards this aim, REBECCA will develop a purely European complete Hardware(HW) and Software(SW) stack around a RISC-V CPU, which will provide significantly higher levels of a)...

cordis.europa.eu

Hi Bravo,

Klepsydra may well turn out to be Mrs Danvers to Akida's Mrs de Winter2 before too much water has flowed under the bridge.

https://klepsydra.com/ai-for-space/

What are the challenges for onboard artificial intelligence?​

There are several challenges for implementing onboard AI:

1. Limited computational resources: Onboard AI systems often operate on devices with limited computing power, such as edge devices or embedded systems. This poses challenges in running complex AI algorithms and models that require significant computational resources. Optimising AI algorithms to work efficiently within these constraints is a challenge.
2. Power consumption: AI algorithms, especially deep learning models, can be computationally intensive and consume significant power. Onboard systems often have limited power sources which require efficient power management strategies to ensure optimal performance without quickly draining the power source.

Real-time processing: Many applications of onboard AI require real-time or near-real-time data processing. The challenge lies in developing AI algorithms and models that can operate within strict latency requirements and make timely decisions within the available processing time.


https://klepsydra.com/klepsydra-ai-technology-evaluation-space-use/

Future missions such as Active Debris Removal will rely on novel high-performance avionics to support image processing and Artificial Intelligence algorithms with large workloads. Similar requirements come from Earth Observation applications, where data processing on-board can be critical in order to provide real-time reliable information to Earth. This new scenario has brought new challenges with it: low determinism, excessive power needs, data losses and large response latency.

In this project, Klepsydra AI is used as a novel approach to on-board artificial intelligence. It provides a very sophisticated threading model combination of pipeline and parallelization techniques applied to deep neural networks, making AI applications much more efficient and reliable. This new approach has been validated with several DNN models and two different computer architectures. The results show that the data processing rate and power saving of the applications increase substantially with respect to standard AI solutions.

https://klepsydra.com/what-is-earth-observation/

What does Klepsydra do regarding Earth Observation?​

Earth Observation (EO) satellites require a suite of applications to perform their mission successfully. The most critical application is data acquisition and processing, which includes collecting and processing data obtained by the satellite’s sensors.

SAR sensors are a type of active remote sensing system that uses radio waves to produce images of the Earth’s surface. On the other hand, Hyperspectral cameras are imaging devices that can capture data across a broad range of electromagnetic wavelengths, including both visible and non-visible parts of the spectrum. Compared to traditional cameras that capture data across only three color channels (red, green, and blue), Hyperspectral cameras can provide a more detailed and
comprehensive view of a scene.

Klepsydra enables the processing of large volumes of data produced by SAR sensors and Hyperspectral cameras used in Earth observation satellites, using artificial intelligence techniques.

Klepsydra is software (they also claim some snazzy models (supra), but it runs on a computer:

https://klepsydra.com/ai-on-the-edge/

 

[7für7]

hmmm maybe you missed the Linked post from their official account

“come on guys.. it was an exhibition where companies could check out our amazing technology.. we took some pictures so what? It’s not our fault you interpret everything aa “akida inside”. != "We welcomed partners like PROPHESEE, Edge Impluse, and RTX..."


View attachment 76123

No to be honest I didn’t missed it. You have to read it differently in terms of the comma placement; then it makes more sense. “And RTX…” – there are two commas, one before and one after. So, the sentences are not directly connected and should be understood separately.

 

[Guzzi62]

I understood to take the dot joining on here with a pinch of salt !

But now cant even rely on whats released on BrainChip comms .... loosing confidence



View attachment 76121

 

[Tothemoon24]

More info in the link​

https://www.embedded.com/neuromorphic-computing-the-new-frontier-of-computer-science-inspired-by-the-human-brain

​

​

Advanced AI for robotics and automotive with BrainChip’s Akida​

The flagship product of BrainChip, another pioneer in neuromorphic computing, is the Akida chip, designed for real-time AI applications, such as robotics, autonomous vehicles and intelligent video surveillance. Akida is based on a spiking neural network technology that mimics the functioning of biological brain, making the chip highly energy-efficient and suitable for edge systems.

One of the distinguishing features of Akida is its ability to learn incrementally, meaning that once implemented in a system, it can improve its performance without the need for complete retraining, a huge advantage for applications such as autonomous driving, in which vehicles must be able to constantly adapt to new situations and environments.

BrainChip has partnered with several companies in the automotive and defense sectors to integrate Akida into AV control systems. The chip has been successfully tested in a variety of applications, including advanced vision systems and radar sensors, showing remarkable performance in terms of processing speed and low power consumption.

Additionally, Akida’s ability to process data in real time makes it particularly suitable for robots that require fast and reliable decisions in dynamic environments.

Prospects and future applications of neuromorphic computing​

The shift from the von Neumann architecture to neuromorphic chips marks a fundamental evolution in the design of modern computing systems. While traditional architecture has provided the foundation, neuromorphic computing chips offer a new computational perspective by mimicking the dynamics of the human brain, enabling efficient and parallel processing. This shift addresses the inherent limitations of the von Neumann architecture and paves the way for new applications and an era of more advanced and adaptable AI.

The potential for neuromorphic computing is enormous and could revolutionize fields such as AI, robotics, automotive and healthcare. Future applications include intelligent medical devices that can monitor and diagnose medical conditions in real time, home robots that interact more naturally with humans, and AVs with highly responsive control systems. Companies such as Qualcomm and BrainChip are demonstrating with real-world cases that this technology is no longer just a theoretical concept but a rapidly evolving reality, with applications that are already revolutionizing various industrial sectors.

One of the strategic objectives for neuromorphic system designers is the integration of this new architecture into traditional workflows. Although several companies are already showing initial successes, large-scale adoption requires a more mature and robust hardware and software infrastructure.


The innovative neuromorphic approach can also revolutionize the way AI systems are developed, reducing energy requirements and increasing processing speed. Continued research in this field could lead to a new generation of devices capable of performing complex cognitive tasks with unprecedented efficiency, redefining the very concept of learning.

 

[Dr E Brown]

I cannot say for sure just pure speculation like everything else.... Tony's statement was NO partnership between RTX and Brainchip.....that does not mean they are not a customer via a different channel, eg Megachips, etc. maybe still an end user. IMO

That is NOT what Tony said. He stated there is no COMMERCIAL partnership agreement. They can be partners in progressing the technology and not have a commercial agreement.