BRN Discussion Ongoing

[JB49]

Joshua Yang Secures Center of Excellence for Work on Neuromorphic Computing​

Landon Hall | January 18, 2024

USC will lead Air Force-funded Center that includes four other universities.​

Joshua Yang, professor of electrical and computer engineering at the USC Ming Hsieh Department of Electrical and Computer Engineering, has secured funding from the U.S. Air Force to create a Center of Excellence based on his work on neuromorphic computing.

(CREDIT: MINGYI RAO AND GLENN GE. MEMRISTOR AI CHIP UNDER EXTREME SPACE ENVIRONMENT. JOSHUA YANG. THIS IMAGE WAS GENERATED WITH THE ASSISTANCE OF AI.)
Yang and USC will lead the center, which also includes researchers from UCLA, Duke University, the University of Texas-San Antonio, and the Rochester Institute of Technology. The entity is called the Center of Neuromorphic Computing and Extreme Environment, or CONCRETE.
A review panel from the Air Force’s Office of Scientific Research and the Air Force Research Laboratory (AFRL) unanimously recommended five years’ worth of funding, based on a proposal titled “Extreme Neuromorphic Materials and Computing.”
Neuromorphic computing is a type of computer engineering modeled after the function of the human brain and nervous system. “A major part of the effort is to build machine-learning or AI accelerators,” Yang said. “They are inspired by the human brain, learn from the brain, and they approach the efficiency of the brain.”
“The brain is the most energy-efficient computer known so far — way better than any supercomputer,” Yang added. “Orders of magnitude better, because we actually consume much less energy. Moreover, the brain can learn much faster and is much more error-tolerant, and much more adaptive to the environment. We tend to learn more and emulate more, so that we can build a much better computing system that’s more sustainable and faster and more robust.”
Sustainability is at the heart of the “Extreme Environment” portion of the CONCRETE acronym. Defense labs, including those at the Air Force, need computer systems that can function under less-than-ideal conditions.
Another pivotal objective of the center is to fortify collaborations between universities and the AFRL, with a focus on educating the workforce in anticipation of future computing requirements for the Department of Defense (DoD).
Securing this esteemed award to establish a center on neuromorphic computing aligns seamlessly with USC’s recent launch of the Frontiers of Computing.

Apparently has 118 granted patents. Hopefully nothing that may obstruct us.

 

[Fact Finder]

Not sure if previously posted

View attachment 54811
View attachment 54812

Hi TLS
I have not seen it before.

Very big claim to make "This collaboration led to the creation of efficient and Embedded Edge Designed solutions, which are now deployed in real-world applications".

This statement would be readily seen by those he works with at Brainchip and his peers in the industry.

If untrue it would be very easily proven to be a lie so on balance unless proven otherwise I would be inclined to believe the accuracy of his claim.

Believing this claim I think I would excluded the VVDN AKIDA Edge Box as it has not even appeared for sale on Brainchip or VVDN's websites so unlikely to have been deployed in real-world applications. I think the same situation would apply to Unigen's AKIDA powered Cupcake which is due Aprilish I think but certainly not available yet. Microchip seems to be out from the language used in the podcast talking about being now able to take it to customers.

Well we are hopefully in for an at least pleasant surprise in due course.

My opinion only DYOR
Fact Finder

 

[Tothemoon24]

Paid Program: The AI-Ready Data Center

Find out how modernizing data centers can help companies be AI-ready, save energy and reduce operating costs.

partners.wsj.com

Business​

Jan 21, 2024

Print Edition

Video

The AI-Ready Data Center​

Companies are racing to invest in AI. Are their data centers ready?
After years of talking about the potential for artificial intelligence to take the world by storm, it’s finally happening, as companies race to incorporate everything from machine learning-driven predictive analytics to generative AI chatbots into their businesses. And the AI boom is already resulting in bottom-line impacts. About 92% of large companies that have invested in AI are seeing returns on their investments, according to a 2022 survey by NewVantage Partners.
However, success in AI can come at a cost, particularly when it comes to upgrading infrastructure such as data centers that were built to serve more modest needs. A recent forecast by global market intelligence firm IDC shows business investments in AI-centric systems will reach $251 billion by 2027, with a compound annual growth rate of 31.4%.
These infrastructure upgrades are critical, as AI workloads place much more intense demand on servers than traditional applications, says Matt Kimball, vice president and principal analyst, data centers, at Moor Insights & Strategy.
Machine learning, for example, can have a big impact on storage, power and compute needs. “AI puts a big strain on a data center that’s been used to run traditional workloads,” he says. “It’s a completely different set of requirements dictating what kind of infrastructure you need.”
Given the much higher demands that AI places on data centers, companies are wise to start looking at how they’ll equip them with the right infrastructure. Those that don’t modernize risk falling behind in a competitive landscape where speed, agility and actionable insights are key.

High Demand​

A dilemma for many companies looking to make AI investments is that data center space is already scarce, says Robert Hormuth, corporate vice president of architecture and strategy in the Data Center Solutions Group at AMD. Companies can’t simply add more racks and consider the job done if there’s no place to put them.
AMD offers a significant performance-per-watt energy efficiency advantage as well as a consolidation advantage. People can now adopt it in their current data centers to help power their AI needs.
Robert Hormuth, Corporate Vice President of Architecture and Strategy, Data Center Solutions Group, AMD
“It’s putting immense pressure on operators, whether on-premises or in colocation centers, to find the space and capacity to bring all these new services online,” he says. “In the past, we always talked about how there was this normal evolution of consolidation, retiring equipment and bringing in more energy efficiency to save cost and power, but now it’s happening at a new level unseen ever before.”
The heightened capacity and computational power demands of AI are creating a race for companies to modernize their data centers. That can mean everything from upgrading servers to higher-performance ones that allow for consolidation—boosting data center capacity—to increasing utilization of their current equipment.
“AMD offers a significant performance-per-watt energy efficiency advantage as well as a consolidation advantage. People can now adopt it in their current data centers to help power their AI needs, rather than spending the huge amount of time and money it would take to build a new data center,” Hormuth says.
Upgrading to servers with the latest generation of CPUs can make a big difference. Today’s 128-core AMD EPYC server processors have two times the computing cores of prior generation processors. This leads to higher levels of performance generation over generation. EPYC processors can help companies gain the compute power they need for AI while also helping them free up data center space.
“The more cores I can stuff into a server, the more I can do with it,” Kimball says. “If I can double my core count on a per-rack basis, I should be able to reduce my footprint significantly.”

Savings for the Future​

It’s important for companies to implement servers with CPUs and GPUs that can facilitate not only the AI models they’re currently planning for, but also those with the compute power and high performance that provide the flexibility to evolve as the models and algorithms progress, Hormuth says. “As the models, algorithms and networking change, having that flexibility to ride the curve is going to be really important,” he adds.
Beyond making way for AI, modernizing the data center also comes with a range of other benefits. In adopting more efficient technologies and strategies, businesses can significantly lower capital expenses and reduce their data center footprint by requiring fewer servers to manage the same workload. This could translate to lower energy consumption and operational costs. “So it’s good from a sustainability perspective, and it’s good from a financial perspective,” Kimball adds.
The 4th Gen AMD EPYC is the right server processor choice as companies look to make their data centers AI-ready, Hormuth says, because it is optimized for higher performance, throughput and energy efficiency.
“If you look at energy efficiency benchmarks in the industry, you will find AMD EPYC is hands down the leader in overall energy efficiency,” he says. “And to deliver that with fewer servers, again drives that consolidation and helps create that space to bring new innovations in your data center.”

 

[Crestman]

Just came across this, doesnt mention Akida but how many actual neuromorphic chips are out there?

Neuromorphic hardware for somatosensory neuroprostheses - PubMed

In individuals with sensory-motor impairments, missing limb functions can be restored using neuroprosthetic devices that directly interface with the nervous system. However, restoring the natural tactile experience through electrical neural stimulation requires complex encoding strategies...

pubmed.ncbi.nlm.nih.gov

In individuals with sensory-motor impairments, missing limb functions can be restored using neuroprosthetic devices that directly interface with the nervous system. However, restoring the natural tactile experience through electrical neural stimulation requires complex encoding strategies. Indeed, they are presently limited in effectively conveying or restoring tactile sensations by bandwidth constraints. Neuromorphic technology, which mimics the natural behavior of neurons and synapses, holds promise for replicating the encoding of natural touch, potentially informing neurostimulation design. In this perspective, we propose that incorporating neuromorphic technologies into neuroprostheses could be an effective approach for developing more natural human-machine interfaces, potentially leading to advancements in device performance, acceptability, and embeddability. We also highlight ongoing challenges and the required actions to facilitate the future integration of these advanced technologies.

 

[Fact Finder]

To clarify, investors who made their fortune or wealth from investing/trading markets, not a successful person who has made wealth elsewhere and invests in the markets..

Such as:
Gil Morales, Mark Minervini, David Ryan, Late William O’Neill, Quint Tatro, John Coates.

In Australia you might like to try Nick Fabrio, Jon Kirk, Tim Walker.. All have presence in either socials or subscription services..

Personal contacts are personal contacts and I wouldn’t give those out publicly..

That’s OK I would not name personal contacts without their permission either.

It’s a problem with making unsupported statements on this anonymous TSEx forum where the one rule is that every unsupported claim should not be given any weight until a poster verifies it for themselves.

It is though very hard to verify the qualifications, opinions and conflicts of interest of the anonymous friend.

My unverified contact is academically, qualified, has worked in the international banking space and currently consults and trades equities, futures and crypto currencies for seven figure returns.

But like your contacts he will need to remain an unverified fiction to which no weight should be attached.

I subscribe to the Warren Buffet and Charlie Munger view of charts which I am sure you know without it being restated. My anonymous contact also holds a similar view but of course that is unverifiable information.

As often is the case this whole discussion has led nowhere because it was never designed to from the very outset.

My opinion only DYOR
Fact Finder

 

[Fact Finder]

Just came across this, doesnt mention Akida but how many actual neuromorphic chips are out there?

Neuromorphic hardware for somatosensory neuroprostheses - PubMed

In individuals with sensory-motor impairments, missing limb functions can be restored using neuroprosthetic devices that directly interface with the nervous system. However, restoring the natural tactile experience through electrical neural stimulation requires complex encoding strategies...

pubmed.ncbi.nlm.nih.gov

In individuals with sensory-motor impairments, missing limb functions can be restored using neuroprosthetic devices that directly interface with the nervous system. However, restoring the natural tactile experience through electrical neural stimulation requires complex encoding strategies. Indeed, they are presently limited in effectively conveying or restoring tactile sensations by bandwidth constraints. Neuromorphic technology, which mimics the natural behavior of neurons and synapses, holds promise for replicating the encoding of natural touch, potentially informing neurostimulation design. In this perspective, we propose that incorporating neuromorphic technologies into neuroprostheses could be an effective approach for developing more natural human-machine interfaces, potentially leading to advancements in device performance, acceptability, and embeddability. We also highlight ongoing challenges and the required actions to facilitate the future integration of these advanced technologies.

Hi C
While it is not commercially available Loihi is used by researchers to demonstrate neuromorphic control of a robotic hand to feel and apply the correct pressure to pick up an object. The demonstration was released last year.

This is one of many links:

https://www.accenture.com/content/d...Precise-and-Energy-efficient-Robotic-Arms.pdf

 

[Fact Finder]

Hi All
I was reading this paper for interest sake and at the very outset the following is stated which fits with statements made by Brainchip about opportunities in the medical diagnostics space:

“Key Messages: The future holds the promise of further advances. Adaptive artificial intelligence will take over diagnostics in manometry and pH impedance testing and patient-driven outcomes may be changed by interactions with artificial intelligence rather than humans. Changes in chip technology will allow higher resolution chips to be carried on smaller devices making extra-esophageal areas where reflux may play a role more accessible to prolonged observation and testing.”

Developments in Gastroesophageal Reflux Disease over the Last 40 Years

Abstract. Background: The last 40 years have seen a remarkable change in our understanding of reflux disease. Summary: These changes encompass disease definition and impact, pathophysiology, diagnostic testing, regulatory oversight of clinical trials, pharmacotherapy, endoscopic, and surgical...

karger.com

My opinion only DYOR
Fact Finder

 

[Crestman]

Hi C
While it is not commercially available Loihi is used by researchers to demonstrate neuromorphic control of a robotic hand to feel and apply the correct pressure to pick up an object. The demonstration was released last year.

This is one of many links:

https://www.accenture.com/content/d...Precise-and-Energy-efficient-Robotic-Arms.pdf

Thanks FF,

Yes I should have said one that isn't going to cost them $10,000.

 

[Rach2512]

Joshua Yang Secures Center of Excellence for Work on Neuromorphic Computing​

Landon Hall | January 18, 2024

USC will lead Air Force-funded Center that includes four other universities.​

Joshua Yang, professor of electrical and computer engineering at the USC Ming Hsieh Department of Electrical and Computer Engineering, has secured funding from the U.S. Air Force to create a Center of Excellence based on his work on neuromorphic computing.

(CREDIT: MINGYI RAO AND GLENN GE. MEMRISTOR AI CHIP UNDER EXTREME SPACE ENVIRONMENT. JOSHUA YANG. THIS IMAGE WAS GENERATED WITH THE ASSISTANCE OF AI.)
Yang and USC will lead the center, which also includes researchers from UCLA, Duke University, the University of Texas-San Antonio, and the Rochester Institute of Technology. The entity is called the Center of Neuromorphic Computing and Extreme Environment, or CONCRETE.
A review panel from the Air Force’s Office of Scientific Research and the Air Force Research Laboratory (AFRL) unanimously recommended five years’ worth of funding, based on a proposal titled “Extreme Neuromorphic Materials and Computing.”
Neuromorphic computing is a type of computer engineering modeled after the function of the human brain and nervous system. “A major part of the effort is to build machine-learning or AI accelerators,” Yang said. “They are inspired by the human brain, learn from the brain, and they approach the efficiency of the brain.”
“The brain is the most energy-efficient computer known so far — way better than any supercomputer,” Yang added. “Orders of magnitude better, because we actually consume much less energy. Moreover, the brain can learn much faster and is much more error-tolerant, and much more adaptive to the environment. We tend to learn more and emulate more, so that we can build a much better computing system that’s more sustainable and faster and more robust.”
Sustainability is at the heart of the “Extreme Environment” portion of the CONCRETE acronym. Defense labs, including those at the Air Force, need computer systems that can function under less-than-ideal conditions.
Another pivotal objective of the center is to fortify collaborations between universities and the AFRL, with a focus on educating the workforce in anticipation of future computing requirements for the Department of Defense (DoD).
Securing this esteemed award to establish a center on neuromorphic computing aligns seamlessly with USC’s recent launch of the Frontiers of Computing.

Yang and USC will lead the center, which also includes researchers from UCLA, Duke University, the University of Texas-San Antonio, and the Rochester Institute of Technology. The entity is called the Center of Neuromorphic Computing and Extreme Environment, or CONCRETE.

BrainChip Adds Rochester Institute of Technology to its University AI Accelerator Program

Discover the collaboration between BrainChip and the Rochester Institute of Technology in the University AI Accelerator Program.

brainchip.com

 

[Bravo]

Be still my beating heart!!!

​

Intel CEO on the three laws of edge computing, and what it means for AI​

BySean Kinney, Editor in Chief
January 15, 2024

During a CES keynote, Intel CEO Pat Gelsinger predicted “at least” a 10-year AI super cycle as automation is brought into different industries

AI everywhere—that’s the new tagline from Intel as it optimizes its silicon for artificial intelligence (AI) workloads, and partners with its growing stable of device makers and software developers to imbue new and existing hardware and applications with AI. During the Consumer Electronics Show (CES)_ last week, CEO Pat Gelsinger touted the consumer and enterprise benefits of the AI PC, and called out more than 100 engagements with AI ISV partners.

At CES Intel announced the Core 14th Gen mobile processor family led by the flagship Intel Core i9-14900HX, and a lineup of Core 14th Gen desktop processors at 65-watt and 35-watt levels to support a wide range of edge devices. These launches were preceded by the December announcement of a set of major AI-focused upgrades on the client and server sides; for data center compute, the 5th Gen Intel Xeon family now has AI accelerate in every core, and the Gaudi 3 AI accelerator is on schedule for a 2024 release.

In conjunction with the December announcements, Gelsinger said, “Intel is on a mission to bring AI everywhere through exceptionally engineered platforms, secure solutions and support for open ecosystems.”


In Las Vegas for a keynote interview at CES, Gelsinger talked through the competitive dynamics at place in the AI PC space as Arm-based competitors push competing products into the markets. The competition, he said, “would help establish the category,” which Intel will then win based on volumes and ecosystems. “I believe this is a defining moment for the PC,” Gelsinger said. “And as [former Intel CEO] Andy Grove said, the PC [is] the ultimate Darwinian device. I think we’re in one of those Cambrium moments for the Darwinian PC.”

Looking ahead, Gelsinger predicted that the coming wave of AI investment and adoption would be at least as durable and significant as the immediate post-internet launch of new ways to drive customer engagement. “I think we’re going to be in a 10-plus-year cycle of figuring it out in different industries.”

Intel led its CES messaging with AI PCs. Asked about how on-device AI adoption would unfold, Gelsinger said adoption of AI-enabled edge devices—phones, PCs, on-premise compute, etc…—would adhere to what he called the three laws of edge computing.

• “First is the laws of economics. It’s cheaper to do it on your device…I’m not renting cloud servers.”
• “Second is the laws of physics. If I have to round-trip the data to the cloud and back, it’s not going to be as responsive as I can do locally.”
• “And third is the laws of the land. Am I going to make my data to the cloud or am I going to keep it on my local device?”

He summarized: “I think those three laws…will drive more of these AI use cases to the devices that we use across the edge, and that’s part of what we mean when we say, ‘AI everywhere.’…I think it’s going to be a thrilling time of new use cases emerging…We’re going to be enjoying this for many CESes to come…We’re just getting started.”

Intel CEO: Three laws of edge computing, and what it means for AI

With new products for client and servers, Intel CEO Pat Gelsinger took to CES to laud the benefits of AI everywhere.

www.rcrwireless.com

 

[Fact Finder]

Thanks FF,

Yes I should have said one that isn't going to cost them $10,000.

Hi C
Thanks to your post I stumbled across this link:

First Ever in Canada: ABR to Showcase New Intel Loihi Chip Running ABR's Reach Robotic Controller this Monday at OCE Discovery 2018 at the MTCC.

Come to the Metro Toronto Convention Centre to OCE's Discovery 2018 Show (https://www.ocediscovery.

www.linkedin.com

If you add the comment by Mike Davies last year that a commercial Loihi may still be another five years it really does put into proper perspective what Brainchip had on display at CES 2024.

Except that industry figures have verified what Brainchip is doing commercially with AKIDA in live demonstrations it boarders on unbelievable.

But I suppose that’s what Science Fiction is something so advanced it is unbelievable based on conventional wisdom.

My opinion only DYOR
Fact Finder

 

[Tothemoon24]

Some interesting upcoming tech talks with ARM for those that maybe interested in registering

Tech Talks

Meet with experts from the Arm ecosystem for tech talks covering best practices and the latest trends across AI, automotive, consumer technology, infrastructure and IoT.

www.arm.com

 

[AusEire]

To clarify, investors who made their fortune or wealth from investing/trading markets, not a successful person who has made wealth elsewhere and invests in the markets..

Such as:
Gil Morales, Mark Minervini, David Ryan, Late William O’Neill, Quint Tatro, John Coates.

In Australia you might like to try Nick Fabrio, Jon Kirk, Tim Walker.. All have presence in either socials or subscription services..

Personal contacts are personal contacts and I wouldn’t give those out publicly..

It took Nick Fabrio 4 years to make 1million while trading.

I mean absolutely no disrespect towards him but I made the same working as a carpenter(that being said my back is absolutely fucked and I doubt his is)

I absolutely cannot stand the next person I'm going to mention (because she's a dishonest shifty operator) and she's currently doing her thing over on HC. But good for her she made a decent buck(far more than Fabrio in half the time) while shitting on all of us. Dolchi(loooochi) is a far superior trader than Fabrio and she donates her time for free on HC. No subscription required.

Btw these people you mentioned are DAY TRADERS. They are not investors. They make up a TEENY TINY MINORITY of people who make money day trading. MOST people that trade lose money. That is a statistical fact. It's incredibly risky.

So again I'll ask you. Who are these REAL INVESTORS you speak of?

 

[AusEire]

My post was a bit of a tongue in cheek dig at the idea I had attended a secret meeting.

When you think about it the risk associated with different groups of unrelated shareholders and management of a company meeting across three capital cities is far less than if an individual shareholder engages in one on one private communications with someone in a company.

Establishing a relationship, perhaps becoming friends through shared experiences then playing on the trust that comes with that friendship to extract intelligence from the employee concerned. Such communications would never come to light unless the shareholder concerned revealed them.

Individual posters particularly on HC constantly claim to have communicated with the company and to have received a totally unsatisfactory reply but how can these anonymous claims ever be verified.

The company would likely never communicate with someone calling themselves Dingo Borat or Fact Finder in a communication so quite hard to link up the claim on HC with an actual communication.

Imagine writing to the company trying to verify this post and asking did you give a totally unsatisfactory reply to a shareholder in the last 14 days. I'm sorry but I can't tell you their name or address only that they call themselves A Dead Loss on HC.

Downramping is such an easy game to play.

My opinion only DYOR
Fact Finder

Over Xmas @Luppo71 and I held a secret Brainchip meeting and we were the only ones to show up. I was quite disappointed as I was expecting Sean and TD to turn up.

 

[goodvibes]

Does somebody knows why Renesas disappeared from partner overview by brainchip? Wasn’t it on integration partners?

Partners

Become a partner!

brainchip.com

 

[AusEire]

Does somebody knows why Renesas disappeared from partner overview by brainchip? Wasn’t it on integration partners?

Partners

Become a partner!

brainchip.com

Correct me if I'm wrong but I'm certain that Renesas were only an IP licensee and were never technology partners.

 

[David holland]

Correct me if I'm wrong but I'm certain that Renesas were only an IP licensee and were never technology partners.

Are Technical partners as good as a IP licence

 

[Pom down under]

Over Xmas @Luppo71 and I held a secret Brainchip meeting and we were the only ones to show up. I was quite disappointed as I was expecting Sean and TD to turn up.

 

[Fact Finder]

Over Xmas @Luppo71 and I held a secret Brainchip meeting and we were the only ones to show up. I was quite disappointed as I was expecting Sean and TD to turn up.

Yep that’s the problem with secret meetings. If you tell people it’s no longer secret but if you don’t tell them they don’t turn up. Personally I quite enjoy my own company so being alone at all my secret meanings has never been an issue.

 

[Diogenese]

Over Xmas @Luppo71 and I held a secret Brainchip meeting and we were the only ones to show up. I was quite disappointed as I was expecting Sean and TD to turn up.

Luppo and AusEire - sounds like The Wolf of O'Connell Street, hey?