BRN Discussion Ongoing
- Thread starter TechGirl
- Start date
Frangipani
Top 20
At least one former BRN employee appears to be of the opinion that Sean Hehir is not the ideal choice for the CEO job (and I strongly suspect he is not the only one):
LinkedIn Login, Sign in | LinkedIn
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.
As I mentioned earlier this month, Anup Vanarse recently moved back to Australia and now works for BrainEye (https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-452508). When I found out he had left BrainChip last year without having another job lined up, only six months after relocating from Perth to California, I suspected personal tensions to be the reason for his departure:
“The fact that he left what looks like a secure job without another one lined up (except for his ongoing side hustle as a remote AI/ML Advisor for NZ-based Scentian Bio, which presumably doesn’t pay the bills) suggests to me he was unhappy in his previous position, possibly due to personal tensions? (Or someone wasn’t happy with him and asked him to leave?)”
(https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-435564)
And in my eyes, a number of his LinkedIn comments/likes since then appear to confirm my suspicion. Of course there will be those who’ll just shrug their shoulders and say nah, it’s nothing, just the voice of a disgruntled ex-employee, but I believe we should listen up when someone who is evidently still highly respected by many BrainChip staff members past and present (look at all the people congratulating him on his new position with BrainEye) gives posts like these a thumbs-up:
"Finn Ryder" as surfer - there's a bit of vocational onomatopoeia.Hi FMF,
Looks like your suspicions were correct!
"Arquimea has deployed Akida with a Prophesee camera on a drone to detect distressed swimmers and surfers in the ocean helping lifeguards scale their services for large beach areas, opting for an event-based computing solution for its superior efficiency and consistently high-quality results."
I wonder if this has anything to do with the recent hiring of Finn Ryder to Development Representative at BrainChip, since he was a Senior Lifeguard and First Responder for the City of Huntington Beach for 5 years prior to joining us?
View attachment 79709
View attachment 79711 View attachment 79710
Brainchip x Prophesee
Enhancing Ocean Safety with Efficient AI-Powered Detection Arquimea has deployed Akida with a Prophesee camera on a drone to detect distressed swimmers and surfers in the ocean helping lifeguards scale their services for large beach areas, opting for an event-based computing solution for its...brainchip.com
Not impressed.. anyway we will not be the only ones. But brain should speed up …just my opinion
... and shark detection.Hi FMF,
Looks like your suspicions were correct!
"Arquimea has deployed Akida with a Prophesee camera on a drone to detect distressed swimmers and surfers in the ocean helping lifeguards scale their services for large beach areas, opting for an event-based computing solution for its superior efficiency and consistently high-quality results."
I wonder if this has anything to do with the recent hiring of Finn Ryder to Development Representative at BrainChip, since he was a Senior Lifeguard and First Responder for the City of Huntington Beach for 5 years prior to joining us?
View attachment 79709
View attachment 79711 View attachment 79710
Brainchip x Prophesee
Enhancing Ocean Safety with Efficient AI-Powered Detection Arquimea has deployed Akida with a Prophesee camera on a drone to detect distressed swimmers and surfers in the ocean helping lifeguards scale their services for large beach areas, opting for an event-based computing solution for its...
rgupta
Regular
Frangipani
Top 20
So with the BrainChip - Prophesee - ARQUIMEA axis of vision confirmed, I guess there’s a good chance, then, that we might have gotten a mention in this November 2024 conference paper I shared on Friday, co-authored by Hélder Rodríguez from ARQUIMEA Research Center? 
After all, we now know that ARQUIMEA’s event-based vision drone solution serves “to detect distressed swimmers and surfers in the ocean helping lifeguards scale their services for large beach areas”.
The link https://portalcientifico.upm.es/en/ipublic/item/10317217 still worked on Friday, but no longer does.
Here is another one:
https://ieeexplore.ieee.org/document/10797412 (abstract only)
After all, we now know that ARQUIMEA’s event-based vision drone solution serves “to detect distressed swimmers and surfers in the ocean helping lifeguards scale their services for large beach areas”.
The link https://portalcientifico.upm.es/en/ipublic/item/10317217 still worked on Friday, but no longer does.
Here is another one:
https://ieeexplore.ieee.org/document/10797412 (abstract only)
Frangipani
Top 20
I’ve always loved this project on the ARQUIMEA website: a real-time detection system of sperm whale blows using thermal imaging. 
Hopefully our technology can one day also help save more of these majestic ocean creatures…
www.arquimea.com
Hopefully our technology can one day also help save more of these majestic ocean creatures…
Artificial Intelligence, AI Orbital | ARQUIMEA
At ARQUIMEA Research Center we believe that Artificial Intelligence is one of the main drivers of transformation in contemporary society.
www.arquimea.com
Maybe this bloke is hard to manage.At least one former BRN employee appears to be of the opinion that Sean Hehir is not the ideal choice for the CEO job (and I strongly suspect he is not the only one):
LinkedIn Login, Sign in | LinkedIn
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.www.linkedin.com
View attachment 79707
As I mentioned earlier this month, Anup Vanarse recently moved back to Australia and now works for BrainEye (https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-452508). When I found out he had left BrainChip last year without having another job lined up, only six months after relocating from Perth to California, I suspected personal tensions to be the reason for his departure:
“The fact that he left what looks like a secure job without another one lined up (except for his ongoing side hustle as a remote AI/ML Advisor for NZ-based Scentian Bio, which presumably doesn’t pay the bills) suggests to me he was unhappy in his previous position, possibly due to personal tensions? (Or someone wasn’t happy with him and asked him to leave?)”
(https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-435564)
And in my eyes, a number of his LinkedIn comments/likes since then appear to confirm my suspicion. Of course there will be those who’ll just shrug their shoulders and say nah, it’s nothing, just the voice of a disgruntled ex-employee, but I believe we should listen up when someone who is evidently still highly respected by many BrainChip staff members past and present (look at all the people congratulating him on his new position with BrainEye) gives posts like these a thumbs-up:
View attachment 79712
View attachment 79714
We used to have people like that working for us and glad they left.
Frangipani
Top 20
Hey @DingoBorat,
here is another humanoid robot video for your collection!
Meet Surgie, the first ever humanoid surgeon, performing direct clinical tasks through teleoperation.
www.linkedin.com
surgie-humanoid.github.io
here is another humanoid robot video for your collection!
Meet Surgie, the first ever humanoid surgeon, performing direct clinical tasks through teleoperation.
Meet Surgie, the first ever humanoid surgeon. | Michael Yip
Meet Surgie, the first ever humanoid surgeon. The UCSD Advanced Robotics and Controls lab has been working hard on developing, for the first time, a humanoid surgeon. Through a combination of teleoperation commands and learned policies, we are able to deploy Surgie into a wide variety of...
Humanoids in Hospitals: A Technical Study of Humanoid Surrogates for Dexterous Medical Interventions
A Technical Study of Humanoid Surrogates for Dexterous Medical Interventions where the experiments were teleoperation of a humanoid robot for medical procedures.
Attachments
DingoBorat
Slim
Hey @DingoBorat,
here is another humanoid robot video for your collection!
Meet Surgie, the first ever humanoid surgeon, performing direct clinical tasks through teleoperation.
Meet Surgie, the first ever humanoid surgeon. | Michael Yip
Meet Surgie, the first ever humanoid surgeon. The UCSD Advanced Robotics and Controls lab has been working hard on developing, for the first time, a humanoid surgeon. Through a combination of teleoperation commands and learned policies, we are able to deploy Surgie into a wide variety of...
View attachment 79730
Humanoids in Hospitals: A Technical Study of Humanoid Surrogates for Dexterous Medical Interventions
A Technical Study of Humanoid Surrogates for Dexterous Medical Interventions where the experiments were teleoperation of a humanoid robot for medical procedures.
View attachment 79732
I reckon I'd rather have a "medical intern" working on me directly, than that shaky thing!
That's only "G1" though..
G3 to G5, most likely won't need the teleoperation and will be smooth as silk..
I thought they were only going to replace boring, mundane and dangerous tasks
..
Tothemoon24
Top 20
Intel Aims For AI Edge As NVIDIA Faces Price Critique
Intel has recently announced a series of initiatives aimed at accelerating the adoption of artificial intelligence (AI) at the edge, marking a significant push…
evrimagaci.org
Intel Aims For AI Edge As NVIDIA Faces Price Critique
Recent developments showcase Intel's strategies amidst pricing concerns over NVIDIA's GPUs and market competition.
Intel has recently announced a series of initiatives aimed at accelerating the adoption of artificial intelligence (AI) at the edge, marking a significant push to simplify the integration of AI with existing infrastructure across various industries, including retail, manufacturing, smart cities, and media. The technology giant unveiled the new Intel AI Edge system, Edge AI suite, and Open Edge platform initiative on March 19, 2025, underlining a commitment to enhancing the efficiency and performance of AI applications deployed in real-world scenarios.Dan Rodriguez, Intel's Corporate Vice President and General Manager of the Edge Computing Group, expressed enthusiasm about the potential for AI integration in existing workflows. "I'm enthusiastic about expanding AI utilization in existing infrastructure and workflows at the edge," Rodriguez stated, highlighting the strong demand for AI-driven solutions that cater to distinct business needs.
According to industry analysts at Gartner, the landscape of data processing is poised for transformation, with predictions indicating that by the end of 2025, fifty percent of enterprise-managed data will be processed outside traditional data centers or clouds. This shift, particularly driven by the integration of AI technologies, is expected to be significant, as companies increasingly rely on data processing at the edge.
Further, it is anticipated that by 2026, at least half of all edge computing deployments will incorporate machine learning, emphasizing the growing importance of AI in data handling and decision-making processes within organizations.
Intel is currently positioned to leverage its extensive footprint in edge deployments; it has over 100,000 real-world edge implementations in collaboration with partners, many of which capitalize on AI functionalities. The new AI technologies are crafted to address multiple industry-specific challenges, underscoring Intel's commitment to enhancing performance standards in edge AI applications.
In a notable development on the following day, March 20, 2025, former Intel CEO Pat Gelsinger issued a biting critique of NVIDIA's pricing structure for its AI GPUs during an interview. He asserted that the current pricing models are "overpriced by 10,000 times the cost required for AI inference," a claim that raises eyebrows around the industry and reflects deep concerns over the affordability of implementing AI solutions.
Gelsinger attributed NVIDIA's recent success to sheer luck rather than a sound strategic framework, suggesting that the company's advancements in AI were more incidental than planned. He emphasized, "AI is in inference, highlighting the need for optimized hardware," pointing to the necessity for improved cost structures as the AI market rapidly evolves.
The discussion surrounding NVIDIA's AI GPU pricing cannot be taken lightly. These GPUs, designed for data center applications, are traded in the tens of thousands of dollars range, making them significantly pricier compared to more affordable specialized hardware developed for inference tasks. Gelsinger's remarks suggest not only industry-wide implications for hardware production but also a serious reassessment of market competitiveness.
Despite Intel’s efforts in the AI domain, the company has faced considerable challenges in maintaining its competitive edge. Recently, it discontinued development of the 'Falcon Shores' AI chip and is now narrowing its focus on the 'Jaguar Shores' initiative. This strategic pivot reflects a recognition of the fierce competition present in the AI semiconductor market, wherein companies like NVIDIA and AMD are currently leading with innovative AI solutions.
Intel’s 'Gaudi' series also aims to deliver cost-effective performance. However, critics argue that its performance falls short when compared to powerhouses like NVIDIA's 'Hopper' and AMD's 'Instinct' lines. This competitive disadvantage is causing Intel to reevaluate its offerings in a landscape that increasingly prioritizes computational efficiency alongside performance metrics.
Looking ahead, Intel is pinning its hopes on the Jaguar Shores line as it seeks to re-establish a foothold in the AI market. However, skepticism remains regarding enterprises’ willingness to pivot away from NVIDIA’s established ecosystem, which is bolstered by its proprietary development environment, CUDA. This ecosystem has proven to be a powerful leverage point, facilitating varying AI applications beyond mere hardware comparisons.
As the industry navigates these turbulent waters, Gelsinger’s statements highlight the urgency for Intel to not only build technologically superior products but also foster a comprehensive ecosystem that includes robust software support and greater cost efficiency. Should the demand for optimized hardware solutions for AI inference grow, as Gelsinger suggests it might, Intel could regain its footing in a rapidly shifting market.
Ultimately, the future of AI market competition appears to be in a constant state of flux, characterized by emerging technologies and changing consumer expectations. The initiatives launched by Intel, coupled with the critical insights shared by industry veterans, make for an intriguing narrative, but the company's ability to adapt and innovate will determine its success in this burgeoning field.
Frangipani
Top 20
Intel Accelerates AI at the Edge Through an Open Ecosystem - Edge AI and Vision Alliance
Intel empowers partners to seamlessly integrate AI into existing infrastructure with its new Intel AI Edge Systems, Edge AI Suites and Open Edge Platform software. What’s New: Intel is unveiling its new Intel® AI Edge Systems, Edge AI Suites and Open Edge Platform initiatives. These offerings...
www.edge-ai-vision.com
Intel’s Edge AI Portfolio
Get help deploying solutions with Intel’s Edge Portfolio, a suite of partner offerings, reference designs, Edge AI Suites, and the Open Edge platform.
Pom down under
Top 20
When is LDA selling? As by the look of things it’s started already and will probably stop once the sell side evens up with the buy side again. Oh the sorters must love LDA so they can close there positions 
Smoothsailing
Regular
Wouldn’t hurt them to yell out Brainchips name in a few paragraphs at the Edge, how long this piece of string.Intel Aims For AI Edge As NVIDIA Faces Price Critique
Intel has recently announced a series of initiatives aimed at accelerating the adoption of artificial intelligence (AI) at the edge, marking a significant push…
Intel Aims For AI Edge As NVIDIA Faces Price Critique
Recent developments showcase Intel's strategies amidst pricing concerns over NVIDIA's GPUs and market competition.
Intel has recently announced a series of initiatives aimed at accelerating the adoption of artificial intelligence (AI) at the edge, marking a significant push to simplify the integration of AI with existing infrastructure across various industries, including retail, manufacturing, smart cities, and media. The technology giant unveiled the new Intel AI Edge system, Edge AI suite, and Open Edge platform initiative on March 19, 2025, underlining a commitment to enhancing the efficiency and performance of AI applications deployed in real-world scenarios.
Dan Rodriguez, Intel's Corporate Vice President and General Manager of the Edge Computing Group, expressed enthusiasm about the potential for AI integration in existing workflows. "I'm enthusiastic about expanding AI utilization in existing infrastructure and workflows at the edge," Rodriguez stated, highlighting the strong demand for AI-driven solutions that cater to distinct business needs.
According to industry analysts at Gartner, the landscape of data processing is poised for transformation, with predictions indicating that by the end of 2025, fifty percent of enterprise-managed data will be processed outside traditional data centers or clouds. This shift, particularly driven by the integration of AI technologies, is expected to be significant, as companies increasingly rely on data processing at the edge.
Further, it is anticipated that by 2026, at least half of all edge computing deployments will incorporate machine learning, emphasizing the growing importance of AI in data handling and decision-making processes within organizations.
Intel is currently positioned to leverage its extensive footprint in edge deployments; it has over 100,000 real-world edge implementations in collaboration with partners, many of which capitalize on AI functionalities. The new AI technologies are crafted to address multiple industry-specific challenges, underscoring Intel's commitment to enhancing performance standards in edge AI applications.
In a notable development on the following day, March 20, 2025, former Intel CEO Pat Gelsinger issued a biting critique of NVIDIA's pricing structure for its AI GPUs during an interview. He asserted that the current pricing models are "overpriced by 10,000 times the cost required for AI inference," a claim that raises eyebrows around the industry and reflects deep concerns over the affordability of implementing AI solutions.
Gelsinger attributed NVIDIA's recent success to sheer luck rather than a sound strategic framework, suggesting that the company's advancements in AI were more incidental than planned. He emphasized, "AI is in inference, highlighting the need for optimized hardware," pointing to the necessity for improved cost structures as the AI market rapidly evolves.
The discussion surrounding NVIDIA's AI GPU pricing cannot be taken lightly. These GPUs, designed for data center applications, are traded in the tens of thousands of dollars range, making them significantly pricier compared to more affordable specialized hardware developed for inference tasks. Gelsinger's remarks suggest not only industry-wide implications for hardware production but also a serious reassessment of market competitiveness.
Despite Intel’s efforts in the AI domain, the company has faced considerable challenges in maintaining its competitive edge. Recently, it discontinued development of the 'Falcon Shores' AI chip and is now narrowing its focus on the 'Jaguar Shores' initiative. This strategic pivot reflects a recognition of the fierce competition present in the AI semiconductor market, wherein companies like NVIDIA and AMD are currently leading with innovative AI solutions.
Intel’s 'Gaudi' series also aims to deliver cost-effective performance. However, critics argue that its performance falls short when compared to powerhouses like NVIDIA's 'Hopper' and AMD's 'Instinct' lines. This competitive disadvantage is causing Intel to reevaluate its offerings in a landscape that increasingly prioritizes computational efficiency alongside performance metrics.
Looking ahead, Intel is pinning its hopes on the Jaguar Shores line as it seeks to re-establish a foothold in the AI market. However, skepticism remains regarding enterprises’ willingness to pivot away from NVIDIA’s established ecosystem, which is bolstered by its proprietary development environment, CUDA. This ecosystem has proven to be a powerful leverage point, facilitating varying AI applications beyond mere hardware comparisons.
As the industry navigates these turbulent waters, Gelsinger’s statements highlight the urgency for Intel to not only build technologically superior products but also foster a comprehensive ecosystem that includes robust software support and greater cost efficiency. Should the demand for optimized hardware solutions for AI inference grow, as Gelsinger suggests it might, Intel could regain its footing in a rapidly shifting market.
Ultimately, the future of AI market competition appears to be in a constant state of flux, characterized by emerging technologies and changing consumer expectations. The initiatives launched by Intel, coupled with the critical insights shared by industry veterans, make for an intriguing narrative, but the company's ability to adapt and innovate will determine its success in this burgeoning field.
Generally bagging an ex employer publicly is risky. An emotionally intelligent person would probably hold himself together and understand its not a good career move.
Bag one employer it may turn off other prospective employers.
He may well have been hard done by but best to be discrete.
I agree that that you give your all at work but its best not to emotionally tie yourself to your employer. Emotionally tied people risk huge personal downers if things go sour.
This statement seems to be affecting us as well.Intel Aims For AI Edge As NVIDIA Faces Price Critique
Intel has recently announced a series of initiatives aimed at accelerating the adoption of artificial intelligence (AI) at the edge, marking a significant push…
Intel Aims For AI Edge As NVIDIA Faces Price Critique
Recent developments showcase Intel's strategies amidst pricing concerns over NVIDIA's GPUs and market competition.
Intel has recently announced a series of initiatives aimed at accelerating the adoption of artificial intelligence (AI) at the edge, marking a significant push to simplify the integration of AI with existing infrastructure across various industries, including retail, manufacturing, smart cities, and media. The technology giant unveiled the new Intel AI Edge system, Edge AI suite, and Open Edge platform initiative on March 19, 2025, underlining a commitment to enhancing the efficiency and performance of AI applications deployed in real-world scenarios.
Dan Rodriguez, Intel's Corporate Vice President and General Manager of the Edge Computing Group, expressed enthusiasm about the potential for AI integration in existing workflows. "I'm enthusiastic about expanding AI utilization in existing infrastructure and workflows at the edge," Rodriguez stated, highlighting the strong demand for AI-driven solutions that cater to distinct business needs.
According to industry analysts at Gartner, the landscape of data processing is poised for transformation, with predictions indicating that by the end of 2025, fifty percent of enterprise-managed data will be processed outside traditional data centers or clouds. This shift, particularly driven by the integration of AI technologies, is expected to be significant, as companies increasingly rely on data processing at the edge.
Further, it is anticipated that by 2026, at least half of all edge computing deployments will incorporate machine learning, emphasizing the growing importance of AI in data handling and decision-making processes within organizations.
Intel is currently positioned to leverage its extensive footprint in edge deployments; it has over 100,000 real-world edge implementations in collaboration with partners, many of which capitalize on AI functionalities. The new AI technologies are crafted to address multiple industry-specific challenges, underscoring Intel's commitment to enhancing performance standards in edge AI applications.
In a notable development on the following day, March 20, 2025, former Intel CEO Pat Gelsinger issued a biting critique of NVIDIA's pricing structure for its AI GPUs during an interview. He asserted that the current pricing models are "overpriced by 10,000 times the cost required for AI inference," a claim that raises eyebrows around the industry and reflects deep concerns over the affordability of implementing AI solutions.
Gelsinger attributed NVIDIA's recent success to sheer luck rather than a sound strategic framework, suggesting that the company's advancements in AI were more incidental than planned. He emphasized, "AI is in inference, highlighting the need for optimized hardware," pointing to the necessity for improved cost structures as the AI market rapidly evolves.
The discussion surrounding NVIDIA's AI GPU pricing cannot be taken lightly. These GPUs, designed for data center applications, are traded in the tens of thousands of dollars range, making them significantly pricier compared to more affordable specialized hardware developed for inference tasks. Gelsinger's remarks suggest not only industry-wide implications for hardware production but also a serious reassessment of market competitiveness.
Despite Intel’s efforts in the AI domain, the company has faced considerable challenges in maintaining its competitive edge. Recently, it discontinued development of the 'Falcon Shores' AI chip and is now narrowing its focus on the 'Jaguar Shores' initiative. This strategic pivot reflects a recognition of the fierce competition present in the AI semiconductor market, wherein companies like NVIDIA and AMD are currently leading with innovative AI solutions.
Intel’s 'Gaudi' series also aims to deliver cost-effective performance. However, critics argue that its performance falls short when compared to powerhouses like NVIDIA's 'Hopper' and AMD's 'Instinct' lines. This competitive disadvantage is causing Intel to reevaluate its offerings in a landscape that increasingly prioritizes computational efficiency alongside performance metrics.
Looking ahead, Intel is pinning its hopes on the Jaguar Shores line as it seeks to re-establish a foothold in the AI market. However, skepticism remains regarding enterprises’ willingness to pivot away from NVIDIA’s established ecosystem, which is bolstered by its proprietary development environment, CUDA. This ecosystem has proven to be a powerful leverage point, facilitating varying AI applications beyond mere hardware comparisons.
As the industry navigates these turbulent waters, Gelsinger’s statements highlight the urgency for Intel to not only build technologically superior products but also foster a comprehensive ecosystem that includes robust software support and greater cost efficiency. Should the demand for optimized hardware solutions for AI inference grow, as Gelsinger suggests it might, Intel could regain its footing in a rapidly shifting market.
Ultimately, the future of AI market competition appears to be in a constant state of flux, characterized by emerging technologies and changing consumer expectations. The initiatives launched by Intel, coupled with the critical insights shared by industry veterans, make for an intriguing narrative, but the company's ability to adapt and innovate will determine its success in this burgeoning field.
" However, skepticism remains regarding enterprises’ willingness to pivot away from NVIDIA’s established ecosystem, which is bolstered by its proprietary development environment, CUDA. This ecosystem has proven to be a powerful leverage point, facilitating varying AI applications beyond mere hardware comparisons."
Its hard to take clients from NVIDIA.
Frangipani
Top 20
Check out this newly launched project called ARCHYTAS (ARCHitectures based on unconventional accelerators for dependable/energY efficienT AI Systems), funded by the European Defence Fund (EDF):
www.linkedin.com
defence-industry-space.ec.europa.eu
ARCHYTAS: A European Defence Fund project on energy-efficient AI | ARCHYTAS (EDF) posted on the topic | LinkedIn
Welcome to the official page of ARCHYTAS (EDF), a European Defence Fund Research project! Over the course of three years ARCHYTAS will research and investigate dedicated hardware architectures for energy-efficient artificial intelligence. The project is funded by the European Union’s European...
European Defence Fund (EDF) - Official Webpage of the European Commission.
European Defence Fund (EDF) - Official Webpage of the European Commission.
Last edited:
Frangipani
Top 20
Check out this newly launched project called ARCHYTAS (ARCHitectures based on unconventional accelerators for dependable/energY efficienT AI Systems), funded by the European Defence Fund (EDF):
ARCHYTAS: A European Defence Fund project on energy-efficient AI | ARCHYTAS (EDF) posted on the topic | LinkedIn
Welcome to the official page of ARCHYTAS (EDF), a European Defence Fund Research project! Over the course of three years ARCHYTAS will research and investigate dedicated hardware architectures for energy-efficient artificial intelligence. The project is funded by the European Union’s European...
View attachment 79763
View attachment 79764
European Defence Fund (EDF) - Official Webpage of the European Commission.
European Defence Fund (EDF) - Official Webpage of the European Commission.
View attachment 79765
View attachment 79766 View attachment 79767 View attachment 79768
One of the ARCHYTAS project partners is Politecnico di Milano, whose neuromorphic researchers Paolo Lunghi and Stefano Silvestrini have experimented with AKD1000 in collaboration with Gabriele Meoni, Dominik Dold, Alexander Hadjiivanov and Dario Izzo from the ESA-ESTEC (European Space Research and Technology Centre) Advanced Concepts Team in Noordwijk, the Netherlands*, as evidenced by the conference paper below, presented at the 75th International Astronautical Congress in October 2024:
*(Gabriele Meoni and Dominik Dold have since left the ACT)
“A preliminary successful demonstration is given for the BrainChip Akida AKD1000 neuromorphic processor. Benchmark SNN models, both latency and rate based, exhibited a minimal loss in accuracy, compared with their ANN coun- terparts, with significantly lower (from −50 % to −80 %) EMAC per inference, making SNN of extreme interest for applications limited in power and energy typical of the space environment, especially considering that an even greater improvement (with respect to standard ANN running on traditional hardware) in energy consumption can be expected with SNN when implemented on actual neuromorphic devices. A research effort is still needed, especially in the search of new architectures and training methods capable to fully exploit SNN peculiarities.”
The work was funded by the European Space Agency (contract number: 4000135881/21/NL/GLC/my) in the framework of the Ariadna research program.
Last edited:
Getupthere
Regular
AI Is Moving Out of the Data Center. How Nvidia Wants to Keep Its Edge.
Similar threads
