BRN Discussion Ongoing

Been previously playing with Loihi and Jetson it appears. First direct neuromorphic research I've seen with Sony other than the Prophesee partnership.

Hopefully gets his hands on Akida at some point :)


Lyes Khacef​

PhD, Senior Research Scientist at Sony | Neuromorphic Computing​

Sony Europe​

Zurich, Zurich, Switzerland​

2K followers 500+ connections​


Experience​

  • Sony Europe Graphic

    Senior Research Scientist​

    Sony Europe

    Oct 2022 - Present10 months
    Zurich, Switzerland
    Neuromorphic computing with event-based communication, asynchronous processing and spike-based computing.


  • Publications​

    • Braille letter reading: A benchmark for spatio-temporal pattern recognition on neuromorphic hardware

      Frontiers in Neuroscience - Neuromorphic Engineering November 11, 2022​

      Spatio-temporal pattern recognition is a fundamental ability of the brain which is required for numerous real-world activities. Recent deep learning approaches have reached outstanding accuracies in such tasks, but their implementation on conventional embedded solutions is still very computationally and energy expensive. Tactile sensing in robotic applications is a representative example where real-time processing and energy efficiency are required. Following a brain-inspired computing approach, we propose a new benchmark for spatio-temporal tactile pattern recognition at the edge through Braille letter reading. We recorded a new Braille letters dataset based on the capacitive tactile sensors of the iCub robot's fingertip. We then investigated the importance of spatial and temporal information as well as the impact of event-based encoding on spike-based computation. Afterward, we trained and compared feedforward and recurrent Spiking Neural Networks (SNNs) offline using Backpropagation Through Time (BPTT) with surrogate gradients, then we deployed them on the Intel Loihi neuromorphic chip for fast and efficient inference. We compared our approach to standard classifiers, in particular to the Long Short-Term Memory (LSTM) deployed on the embedded NVIDIA Jetson GPU, in terms of classification accuracy, power, and energy consumption together with computational delay. Our results show that the LSTM reaches ~97% of accuracy, outperforming the recurrent SNN by ~17% when using continuous frame-based data instead of event-based inputs. However, the recurrent SNN on Loihi with event-based inputs is ~500 times more energy-efficient than the LSTM on Jetson, requiring a total power of only ~30 mW. This work proposes a new benchmark for tactile sensing and highlights the challenges and opportunities of event-based encoding, neuromorphic hardware, and spike-based computing for spatio-temporal pattern recognition at the edge.
 
  • Like
  • Fire
  • Love
Reactions: 27 users

mcm

Regular
I've been biding my time in the back ground for a while however feel compelled to say the posts on here over the past few days have been very very revealing. Sincere thanks to all the posters who have uncovered some excellent information. I get the feeling BRN is about to embark on a sustained run upwards. Can't see the SP remaining in the 30's for much longer. When the shorters start covering ... there will be a scramble for BRN shares. It's why I've gorged myself recently on sub-40c shares. Could we be seeing the start of the hockey stick? I think so.
 
  • Like
  • Love
  • Fire
Reactions: 64 users

Evermont

Stealth Mode
Good example below of the competitive environment BrainChip encounter when trying to attract high-caliber AI talent. Might also answer a few queries into why our current remuneration policies must include structure for performance bonuses and other incentives.

Cheers all.

1690424724699.png
 
  • Like
  • Wow
  • Fire
Reactions: 26 users

Cardpro

Regular
I guess we will need to wait till Q3 for them to land new IP contracts for Gen 2 Akidaaaaaaaa
 
  • Like
Reactions: 4 users
Was looking into Rad-hard again and came across this article and within it, the below.

Anyone know much on Coherent Logix?

Only because they apparently use digital neuromorphic and on a skim, can't see if it's something like Loihi driven or who? :unsure:


The quest for commercial-performance electronics in space​

June 21, 2023
Satellite and spacecraft designers seek to apply state-of-the art capabilities of commercial electronics components for short- and long-duration missions in the radiation environments of space.
John Keller



Rad-hard solutions​

Coherent Logix Inc. in Austin, Texas, offers the HyperX family of radiation-hardened software-defined microprocessors. One of the family's early models, the hx2100, has seen design-ins in a variety of space projects, says Michael Doerr, Co-CEO and CTO at Coherent Logix.

"Satellite payloads, communications, or critical control systems in a spacecraft historically have been built stove-piped with one function and one use case," Doerr explains. "That's not the case anymore. Satellites are multi mission capable, and evolve their missions over time, and need reliability. "Today you Need a fully programmable solution up there. Traditional solution is FPGAs, but are high power, and are not friendly to reprogram from a mission standpoint. To build an ASIC would be very expensive. Our solution provides the performance of an ASIC or FPGA, but has the power consumption of an ASIC. Now getting the best of both worlds. Used to be a pure hardware solution and now is a software solution.


The Coherent Logix HyperX family of programmable space processors offer "a fully software-defined platform for communications, video imaging, AI, and adaptive processing on the satellite, with software-defined networking, and active cyber security. Now you have the ability to do that."

The HyperX family offers the ability to do on-orbit processing at data plane rates," Doerr says. "It's an absolute game-changer with what we can do with cyber security and software-defined networking. The HyperX family, he says is a different type of fabric. "Some call it polymorphic, others call it multiprocessor fabric. It's a blend of all of those. Its heritage is looking at biologically inspired systems of how our brain works. It is digital neuromorphic processing -- a unique architecture that is now being accepted into the marketplace."

HyperX can be designed for low-, medium-, geosynchronous orbits -- for short- and long-duration missions, Doerr says. The company's latest offering is the radiation-hardened HyperX: Midnight system on chip (SoC) processor for commercial space applications like communications satellites. HyperX: Midnight offers as much as four times the computing throughput, half the power consumption, and a 40 percent lower price compared to leading radiation-hardened field-programmable gate arrays (FPGAs), company officials say.

Software-defined space processors​

This enables greater capabilities to be packaged in smaller satellite bus sizes to help lower launch costs and accelerating industry growth. HyperX: Midnight also brings software-defined operations, C-programmability, and simplified software debugging to commercial space. Combined with Coherent Logix’s HyperX Studio design suite, these features offer full on-orbit re-programmability, so improvements can continue long after launch. HyperX SoCs support market solutions across several sectors, including consumer electronics, 5G mobile infrastructure, video/broadcast infrastructure, machine vision, sensor processing, and artificial intelligence.
 
  • Like
  • Fire
  • Thinking
Reactions: 28 users

equanimous

Norse clairvoyant shapeshifter goddess
Screenshot_20230727_131047_Brave.jpg
Hi @Beebo

As Rob T would say..."that's a great question" :LOL:

Def an interesting statement for mine.

So, is Nandan talking about pay tv operators / streaming services at the source or is he talking about the end user and things like set top boxes, sticks etc or players in between?

Obviously have pay tv operators like Foxtel, Apple, Disney, Fetch, Stan, Paramount an on and on.

Even though smart TVs these days have streaming apps preloaded or available to DL there is still the standalone receiver space like Chromecast, Android TV, Amazon Fire Stick, NVIDIA Shield and generic Foxtel, Telstra and Fetch STBs.

Very quick article below (not read the white paper) that touches on the tech point on the STBs. Obviously something the industry is looking at.

So it's hard to say where in the chain we are working, though Nandan says "we are working on it". Does this tie in under the telecommunications space we are working with?

Given the margin model statement we know they need a low cost solution but it would potentially provide huge volumes for us.

And what are these "new capabilities" that need promoting? Obviously AI based, so maybe HMI, voice, eye tracking, hand gestures for end users. Who are we partnered with that could or does work in that space maybe?


How Can Pay TV Operators Grow and Thrive in A Post-Pandemic World​

6411b2b41ce94f87e6a4e945_Pay-TV-operators-Main-banner.png


Bleuenn Le Goffic​

VP, Strategy & Business Development
February 23, 2023

1 min
Being an operator in the 21st century has truly been a transformative experience.

The UX expectations from end users are ever increasing as is the number of available services from which they can access and stream content. We’re also seeing traditional Pay TV subscriptions continuing to fall, and according to analysis, this decline is likely to accelerate in 2023. As a result, Pay TV operators face challenging decisions over many crucial factors including the technology choice powering their STBs, how to aggregate content across services, and what business model they should adopt for monetizing their content.

With users becoming more accustomed to churning in and out of subscriptions, it is essential that operators try to stay ahead of the curve and work proactively to both attract and retain users. Will the growth of FAST channels see a decline in the need for large subscription bundles which have been the mainstay of operator propositions? What is the best way forward for an operator? It’s clear that new technology as well as agile work methods need to be accepted and embraced in order to compete in the future.

In this whitepaper, we will walk through some of the key strategic choices you can make as an operator. As always, there is no one-size-fits-all solution and you will need to assess your company and its unique opportunities and challenges before concluding which strategy is right for you. If your organization is big with a large budget, you might be contemplating the acquisition of a media house or buying the exclusive content rights for a major sports league. If size is not in your favor, you will need to come up with different ways to compete.
The first thing that comes to mind is that TV remotes will be a thing of the past. Voice and hand gesture controls will replace it and this is where BRN tech will come in providing effeciency and security.
 
  • Like
  • Love
  • Fire
Reactions: 29 users
Was looking into Rad-hard again and came across this article and within it, the below.

Anyone know much on Coherent Logix?

Only because they apparently use digital neuromorphic and on a skim, can't see if it's something like Loihi driven or who? :unsure:


The quest for commercial-performance electronics in space​

June 21, 2023
Satellite and spacecraft designers seek to apply state-of-the art capabilities of commercial electronics components for short- and long-duration missions in the radiation environments of space.
John Keller



Rad-hard solutions​

Coherent Logix Inc. in Austin, Texas, offers the HyperX family of radiation-hardened software-defined microprocessors. One of the family's early models, the hx2100, has seen design-ins in a variety of space projects, says Michael Doerr, Co-CEO and CTO at Coherent Logix.

"Satellite payloads, communications, or critical control systems in a spacecraft historically have been built stove-piped with one function and one use case," Doerr explains. "That's not the case anymore. Satellites are multi mission capable, and evolve their missions over time, and need reliability. "Today you Need a fully programmable solution up there. Traditional solution is FPGAs, but are high power, and are not friendly to reprogram from a mission standpoint. To build an ASIC would be very expensive. Our solution provides the performance of an ASIC or FPGA, but has the power consumption of an ASIC. Now getting the best of both worlds. Used to be a pure hardware solution and now is a software solution.


The Coherent Logix HyperX family of programmable space processors offer "a fully software-defined platform for communications, video imaging, AI, and adaptive processing on the satellite, with software-defined networking, and active cyber security. Now you have the ability to do that."

The HyperX family offers the ability to do on-orbit processing at data plane rates," Doerr says. "It's an absolute game-changer with what we can do with cyber security and software-defined networking. The HyperX family, he says is a different type of fabric. "Some call it polymorphic, others call it multiprocessor fabric. It's a blend of all of those. Its heritage is looking at biologically inspired systems of how our brain works. It is digital neuromorphic processing -- a unique architecture that is now being accepted into the marketplace."

HyperX can be designed for low-, medium-, geosynchronous orbits -- for short- and long-duration missions, Doerr says. The company's latest offering is the radiation-hardened HyperX: Midnight system on chip (SoC) processor for commercial space applications like communications satellites. HyperX: Midnight offers as much as four times the computing throughput, half the power consumption, and a 40 percent lower price compared to leading radiation-hardened field-programmable gate arrays (FPGAs), company officials say.

Software-defined space processors​

This enables greater capabilities to be packaged in smaller satellite bus sizes to help lower launch costs and accelerating industry growth. HyperX: Midnight also brings software-defined operations, C-programmability, and simplified software debugging to commercial space. Combined with Coherent Logix’s HyperX Studio design suite, these features offer full on-orbit re-programmability, so improvements can continue long after launch. HyperX SoCs support market solutions across several sectors, including consumer electronics, 5G mobile infrastructure, video/broadcast infrastructure, machine vision, sensor processing, and artificial intelligence.
Just another mention / insight into them in an article around their joining the NFP Greening of Streaming around same time as Microsoft.

My question is does Coherent design and manufacture the hardware for neuromorphic software to run on or a total package?



23 February 2023

Microsoft joins GoS, taking shine off novel Coherent Logix​


By Tommy Flanagan

It’s the big one we’ve all been waiting for. Greening of Streaming has revealed its first hyperscaler member – more than six months since first teasing that one of the big three cloud giants was sniffing around the non-profit initiative.

Microsoft is now officially on board, lending its insane scale to Greening of Streaming working groups which are striving for joined-up thinking around the engineering of more sustainable streaming technologies.

With this scale comes a web of insights into where energy demands are taking place across Microsoft’s global technology infrastructure, namely Azure cloud infrastructure. Making that kind of data readily available to Greening of Streaming members should add invaluable color to some of the ongoing data projects, and in many cases accelerate them.

Microsoft wants to educate the world about operating streaming infrastructure at scale, but the giant itself could learn a thing or two from existing Greening of Streaming members. In our experience, the bigger they come, the more resistant they are to change.

Microsoft claims to be championing change from developer to consumer, although it leaves evidence of that to your imagination. There is no indication in the announcement to which GoS working groups Microsoft plans to engage with. However, the timing could be crucial, or simply coincidental, with Greening of Streaming just two weeks ago announcing its LESS (Low Energy Sustainable Streaming) Accord.

One of the early aims of LESS is to gauge industry responses to the idea of encoding video based on energy consumption, rather than on quality. Once the pitchfork-bearing mob subsides, Greening of Streaming plans to produce some kind of palatable framework and then conduct field tests for energy-optimized encoding later this year.

We recall an interaction with Microsoft’s Simon Crownshaw, Worldwide Lead for Media and Entertainment, last year. At the time, Faultline was on the hunt for evidence of Microsoft Azure’s claims about cross-cloud collaboration involving the sharing of sustainability data.

Crownshaw explained that Microsoft has engaged with both competitors and partners as part of industry events and programs to demonstrate the power of the cloud to decarbonize workflows.

No names were mentioned, but Crownshaw did cite that increasing transparency across the industry is paramount in how Microsoft shares data and best practices.

“From a Microsoft perspective, we want to enable industry leaders at the highest level to make the right choices. We are engaged in projects across both customers and potential customers to look at how sustainability will create new business models in the future,” Crownshaw told Faultline.

We are fishing around for more up-to-date comments from Crownshaw and those Microsoft teams engaging directly with Greening of Streaming working groups, and will provide an update once Microsoft is bedded in as an official member.

The immediate downside of Microsoft’s signature is that its giant gravitational pull has drawn the spotlight almost entirely away from another new Greening of Streaming membership, with Coherent Logix joining at the same time.

We are reliably informed that Texas-based Coherent Logix has developed a novel and innovative approach to CPU architecture.

Now in its fourth-generation of HyperX Midnight software-defined SoCs, Coherent Logix has designed the technology with power efficiency top of mind, along with high-performance and upgradability.

These are described as digital neuromorphic semiconductors.

Neuromorphic computing is an approach inspired by the neurons of biological brains, in that compute nodes take the structure of artificial synapses in the brain to process and store data just like neurons.

Faultline first came across the concept of neuromorphic silicon back in 2016, when it wasn’t a concept at all but the Lawrence Livermore National Laboratory in the US had just bought the first neuromorphic chip from IBM Research.

The chip was being sold as TrueNorth at the time, in a deal involving a configuration the equivalent of 16 million neurons and 4 billion synapses, while only consuming the energy equivalent of a tablet.

Even then, this IBM project was eight years in the making – the result of 200 collaborators working together since 2008. IBM boasted that its TrueNorth neuromorphic silicon could be installed just as easily in a smartphone as a supercomputer in the cloud, and that it planned to include these capabilities in its Watson AI program.

We waited another five years until neuromorphic silicon reappeared on our radar, not from IBC but from audio and imaging sensing technology developer DTS, part of Xperi and a sister company to TiVo. DTS announced in April 2021 what claimed to be the world’s first neuromorphic driver monitoring system, designed for gaze tracking, head pose monitoring, identification, and eyelid opening.

Circling back to Coherent Logix, the company claims to have pushed back the breakdown of Moore’s Law by over 20 years. The law states that the number of transistors on a chip doubles approximately every two years, though the cost of computers is halved.

Coherent Logix’s HyperX silicon is targeted at always-on devices. With more and more always-on devices shipping, there are concerns about the rising impact on energy consumption of the connected home. While nominal in the grand scheme of things today, it is inevitable that CPE and connected devices will account for a rising share of global greenhouse gas emissions.

Coherent Logix’s
HyperX processors are therefore targeted at always-on devices, which – just like the human brain – contains processing elements that only consume power when they are thinking.
HyperX processors claim to offer up to 4x the performance of competitors, while consuming just half the power.

With those kind of claims, imagine the kind of impact Microsoft could have on reducing the power consumption of its IT infrastructure and that of its partner network by taking a license with Coherent Logix. Well, it’s just as well the two companies will now be squeezed into the same room.
 
  • Like
  • Fire
  • Thinking
Reactions: 15 users
Looks like they play in the telecommunications/ tv / broadcasting market too.



Coherent Logix Showcases Powerful Hardware and Software-Defined Convergence to Broadcast News and Entertainment Industries at NAB Show

The pioneering company’s new models decrease processing waste and meet consumer demand for green computing while delivering ATSC 3.0 Broadcasts
500NewsWire
NEWS RELEASE BY 500NEWSWIRE
Download PDFDownload multimedia
facebook icon linkedin icon twitter icon pinterest icon email icon Austin, TX | April 14, 2023 08:00 AM Eastern Daylight Time



vczTypz7.jpeg

Austin, TX., April 14, 2023 (500NewsWire) Coherent Logix™, a world leader in high-performance, low power Systems-on-Chips (SoCs) and C-programmable processors for the embedded systems market, will be at the 2023 National Association of Broadcasters (NAB) Show held April 15 - 19 at the Las Vegas Convention Center, located at the ATSC 3.0 Pavilion (W3443).

Coherent Logix’ HyperX family of SoCs includes processors for terrestrial market solutions across several sectors, including consumer electronics, 5G mobile infrastructure, video and broadcast infrastructure, machine vision, sensor processing and production-oriented AI. The company’s new models decrease processing waste and meet consumer demand for green computing.

“Our technology provides flexible configurations of the physical layer for both manufacturers and broadcasters, to enable various operating modes and allow for future technology adaption,” said Joe Fabiano, Coherent Logix Chief Marketing Officer.

Interviews with Joe Fabiano can be arranged by contacting Lesley Gamwell at lgamwell@rountreegroup.com or 404-309-6915.

From early on, Coherent Logix has served as an architect and integral team member of the Advance Television Systems Committee’s (ATSC) standards committee and developing voluntary standards for digital television broadcasting that helped to facilitate interoperability with other media.

“Today, with Coherent Logix technology, design engineers are no longer tethered to the traditional limitations of component design, and instead can build ‘best of breed’ ATSC 3.0 algorithms that fulfill today’s features and tomorrow’s innovations,” said Fabiano.

About Coherent Logix Inc.

Headquartered in Austin, TX, Coherent Logix is a world leader in low-power, high-performance, C-programmable processors for the embedded systems market. They are the inventors of the HyperX architecture, HyperX System-on-a-Chip (SoC) and the HyperX Platform and recently introduced the world’s most advanced System-on-a-Chip for space applications - The HyperX Midnight. The company’s comprehensive portfolio of solutions includes processors, integrated system development tools, optimized libraries, system reference designs, and a customizable system development platform that reduces development complexity and time to market. To learn more, visit https://coherentlogix.com or follow the company on LinkedIn.
 
  • Like
  • Love
  • Fire
Reactions: 14 users

Getupthere

Regular
Volume starting to dry up.

There’s only so much you can buy and sell from the right hand to the left

They will eventually try and push the price up to then sell down to hit stop losses.

Times a ticking

All I’m my opinion.
 
  • Like
  • Fire
Reactions: 27 users

IloveLamp

Top 20
Volume starting to dry up.

There’s only so much you can buy and sell from the right hand to the left

They will eventually try and push the price up to then sell down to hit stop losses.

Times a ticking

All I’m my opinion.
1fq3.gif
 
  • Haha
  • Like
Reactions: 22 users

M_C

Founding Member
  • Like
  • Fire
  • Love
Reactions: 18 users
I knew it. Zeebot secretely testing AKIDA on the forum, or perhaps the first use case..... :)

AI transformers on the edge predicting user likes based on one shot learning of user behaviour on posts that havent even been posted!

1690447789566.png
 
Last edited:
  • Haha
  • Like
Reactions: 18 users

equanimous

Norse clairvoyant shapeshifter goddess
  • Haha
  • Like
Reactions: 15 users

Tothemoon24

Top 20
  • Like
  • Fire
Reactions: 7 users

Tothemoon24

Top 20
  • Like
  • Fire
Reactions: 9 users

miaeffect

Oat latte lover
  • Like
  • Fire
Reactions: 8 users
Wasn't there something not long ago on this or being liked by Rob or someone?

Seems our S&BD Dir also likes it :unsure:


Screenshot_2023-07-27-18-26-43-11_4641ebc0df1485bf6b47ebd018b5ee76.jpg
Screenshot_2023-07-27-18-27-02-83_4641ebc0df1485bf6b47ebd018b5ee76.jpg
 
  • Like
  • Fire
  • Thinking
Reactions: 22 users

M_C

Founding Member
Screenshot_20230727_204625_LinkedIn.jpg
 
  • Like
Reactions: 14 users
Whilst I was checking LinkedIn with Jen's.....quite liking this employee....just saying :)

This sort of quality no doubt supports recent developments of Akida and PVDM, Anil and the team very well.


Olivier Coenen​

Senior Research Scientist, Brainchip​

BrainChip University of California, San Diego​

San Diego County, California, United States
2K followers 500+ connections​



About​

Olivier Coenen has been an early pioneer of brain-inspired processing and modeling with applications to robotics and autonomous systems.

Olivier has led the implementation of event-based methods and algorithms for event-based vision sensors, including fast recognition short-latency object and action recognition using deep event-based networks and spiking networks, event-based: object tracking, stereo vision, optic flow, IMU-driven image stabilization, online frequency estimation, image registration, noise removal, resolution upscaler, image reconstruction from vision data. These have been applied on flying drones, land vehicles, in the retail domain, surveillance & security and others.

Olivier adapted deep networks, such as Transformers, to image processing and context-dependent recognition to retail applications, with implementations in Tensorflow, PyTorch, network compilers and optimizers for accelerator hardware.

At Brain Corp, Olivier developed neural network "models of the motor control system" and designed "artificial nervous systems for UAVs" (small commercial drones). His work directly contributed to closing Series B with Qualcomm Ventures & generated tens of patents.

Olivier established the Neuroscience Group at the Sony Computer Science Laboratory, Paris, which developed technologies in signal processing, machine learning and robotic control inspired from brain processing.

He created and led 6 funding grants, with 2 major projects totaling over 12 million dollars in funding and received 16 academic awards and fellowships.

Olivier has the ability to bring together knowledge from many different sources, identify novel directions, anticipate problems well in advance, establish connections btw apparently disparate areas is one his strengths, which brings in novel perspectives.

In 2004, for example, Olivier created the project SENSOPAC to evolve the state-of-the-art in robotic touch and manipulation, from tactile processing to sensor designs. Clearly, humans and other primates attained tactile information very differently than engineers' approaches. By bringing together roboticists and touch physiologists, the resulting awakening led to the development of one of the most advanced tactile sensors for robot manipulation existing today, the BioTac of SynTouch, LLC.

Olivier was an early adopter at UCSD of artificial neural networks, or parallel distributed processing (PDP) before they became known as deep learning with pioneers such as Terry Sejnowski, Geoff Hinton, D. Rumelhart, R. Hecht-Nielsen, M. Jordan, G. Cottrell, and many others.
 
  • Like
  • Fire
  • Love
Reactions: 40 users
Top Bottom