BRN Discussion Ongoing

[TECH]

Well they are using an alias, but this looks like DeepSeek's patent:

CN118798303A Large language model training method, question and answer method, equipment, medium and productPatent Translate 20240913

ALIYUN FEITIAN HANGZHOU CLOUD COMPUTING TECH CO LTD

Inventors FENG WENFENG; ZHANG YUEWEI; ZENG ZHENYU

The invention provides a large language model training method, a question and answer method, equipment, a medium and a product, and relates to the technical field of artificial intelligence, the training method comprises the following steps: obtaining long text training data, the sequence length of the long text training data being greater than the maximum length of an input text sequence of a pre-trained large language model; increasing a rotation angle base number of a rotation position code of the pre-trained large language model to obtain a modified pre-trained large language model; and training the modified pre-trained large language model by using the long text training data to obtain a trained large language model. In the embodiment, the pre-trained large language model is trained by acquiring the long text training data and increasing the base number of the rotation angles of the rotation position codes, so that the length of the input text sequence is amplified, and the trained large language model can process the long text sequence; and the answer integrity and accuracy of the large language model on questions dependent on long texts and multi-document comparison are improved.

I'll never forget what Peter told me a few years ago when some of the Brainchip team were visiting China, don't leave your laptops lying around or connect through any of their networks (servers)...such trust.

Yes, I think I'm right in saying that we have 2 patents issued in China, we have been assured the process we followed would insure that any domestic (Chinese) company who attempted to steal our trade secrets would be dealt with by the CCP, that I believe came from the Chinese Patent Attorney based in Los Angeles that Brainchip engaged.

On a personal note...I wouldn't trust a Chinese company if my life depended on it, and NO I'm not a racist, I have taught many Chinese how to play golf and individually they are very nice people, it's the control the CCP has over their citizens worldwide, that's the problem...sleeper cells cover the globe, wake up to the fact that ants work as a team once the consequences are laid bare.

Have a nice evening.

 

[Smoothsailing]

I'll never forget what Peter told me a few years ago when some of the Brainchip team were visiting China, don't leave your laptops lying around or connect through any of their networks (servers)...such trust.

Yes, I think I'm right in saying that we have 2 patents issued in China, we have been assured the process we followed would insure that any domestic (Chinese) company who attempted to steal our trade secrets would be dealt with by the CCP, that I believe came from the Chinese Patent Attorney based in Los Angeles that Brainchip engaged.

On a personal note...I wouldn't trust a Chinese company if my life depended on it, and NO I'm not a racist, I have taught many Chinese how to play golf and individually they are very nice people, it's the control the CCP has over their citizens worldwide, that's the problem...sleeper cells cover the globe, wake up to the fact that ants work as a team once the consequences are laid bare.

Have a nice evening.

I was told by a Chinese man himself while working in China that to win is the most important thing to them even if it’s against the initial agreement. I have had several products developed in China over the years and after receiving the samples I went ahead with large orders only to receive a different product. When questioned why there was a difference to the sample their response is this first one sample. I have heard of many cases over the years in many different businesses that receive the same outcome.

 

[Mt09]

I'll never forget what Peter told me a few years ago when some of the Brainchip team were visiting China, don't leave your laptops lying around or connect through any of their networks (servers)...such trust.

Yes, I think I'm right in saying that we have 2 patents issued in China, we have been assured the process we followed would insure that any domestic (Chinese) company who attempted to steal our trade secrets would be dealt with by the CCP, that I believe came from the Chinese Patent Attorney based in Los Angeles that Brainchip engaged.

On a personal note...I wouldn't trust a Chinese company if my life depended on it, and NO I'm not a racist, I have taught many Chinese how to play golf and individually they are very nice people, it's the control the CCP has over their citizens worldwide, that's the problem...sleeper cells cover the globe, wake up to the fact that ants work as a team once the consequences are laid bare.

Have a nice evening.

We had an Inivation centre in China circa March 2020.. maybe it hadn’t quite come online before covid hit..

https://brainchip.com/wp-content/uploads/2020/03/2046907.pdf

 

[CHIPS]

An IT association based in the Ukrainian city of Chernivtsi organised the AI Hills Conference that took place there on Saturday and brought together over 200 participants interested in AI.

Since one of the invited speakers happened to be a gentleman presenting on neuromorphic chips, I decided to have a closer look at the uploaded conference pictures and video, since we’ve known for over a year that Ukraine-based Data Science UA has been playing around with Akida. And bingo - the speaker did indeed represent Data Science UA!



View attachment 76893



View attachment 76894

AI Hills Conference . Chernivtsi IT Cluster

AI Hills Conference

www.it-cluster.cv.ua

View attachment 76889

03 – Google Drive

View attachment 76890

View attachment 76891



View attachment 76892

According to my Google Lens translation, Ilya Babichev started out talking about the advantages and disadvantages of neuromorphic chips in general - the disadvantages being that precision-wise, they have yet to prove themselves more accurate than the “classic” (= von Neumann) architecture, that the software for NC has not yet caught up with the hardware and that well-defined performance benchmarks were still missing.

After introducing AKD1000, he listed at least three different use cases, namely drone detection (obviously a very important real life use case in 2025 Ukraine!), keyword detection and recognition of surface types, such as identifying areas of soil suitable for planting crops with the help of a quadcopter, in areas where forest fires and deforestation have resulted in land that can be repurposed for agriculture.

Wow, great find, Frangipani!

 

[TECH]

We had an Inivation centre in China circa March 2020.. maybe it hadn’t quite come online before covid hit..

https://brainchip.com/wp-content/uploads/2020/03/2046907.pdf

View attachment 76905

The plan changed...in my opinion it was a very wise move...no innovation center, no Brainchip local agent and no distribution center...funny fact...my father was born in Shanghai !

 

[CHIPS]

Can somebody please email Mr Dawe and ask for the top 20 that normally accompanies the quarterly reports .

Thanks .

Why don't you do it yourself? Are we your servants?

 

[IloveLamp]

https://www.embedded.com/brainchip-enables-quantum-venturas-cyberthreat-intelligence-tool-with-low-power-neuromorphic-technology

RunTime - Engineering Talent Solutions on LinkedIn: #neuromorphiccomputing #cybersecurity #edgeai #innovation

Hi folks! 🚀 BrainChip and Quantum Ventura are revolutionizing cybersecurity with neuromorphic tech! Here’s why this is a game-changer: • 𝗔𝗸𝗶𝗱𝗮™…

www.linkedin.com

 

[rgupta]

I'll never forget what Peter told me a few years ago when some of the Brainchip team were visiting China, don't leave your laptops lying around or connect through any of their networks (servers)...such trust.

Yes, I think I'm right in saying that we have 2 patents issued in China, we have been assured the process we followed would insure that any domestic (Chinese) company who attempted to steal our trade secrets would be dealt with by the CCP, that I believe came from the Chinese Patent Attorney based in Los Angeles that Brainchip engaged.

On a personal note...I wouldn't trust a Chinese company if my life depended on it, and NO I'm not a racist, I have taught many Chinese how to play golf and individually they are very nice people, it's the control the CCP has over their citizens worldwide, that's the problem...sleeper cells cover the globe, wake up to the fact that ants work as a team once the consequences are laid bare.

Have a nice evening.

When it comes to power no one better than other. e.g tiktok take the same data whatsoever facebook, whatsapp, Instragram, twitter take from us, the only difference is that share that data with US govt while tiktok do the same with ccp.
Right now china is leader in renewables, battery, mining etc and they are far ahead of western world. In AI they are catching up fast with western world.
I am neither for china or against China but matter of fact everyone salute the rising sun. If you donot know how to adjust with that, that will be a negative.
Dyor

 

[manny100]

Allegations that associates of DeepSeek have made off with a chunk of Open AI and Microsoft data they were not entitled to.
Maybe QV/BRN M/Lockheed-Martin could have closed the entry points presenting opportunities for unauthorised access.

 

[Diogenese]

Allegations that associates of DeepSeek have made off with a chunk of Open AI and Microsoft data they were not entitled to.
Maybe QV/BRN M/Lockheed-Martin could have closed the entry points presenting opportunities for unauthorised access.

DS acknowledge using public domain models.

 

[Frangipani]

A video going along with that paper was uploaded to YouTube yesterday:



Both paper and video relate to another paper and video published by the same Uni Tübingen authors earlier this year. At a cursory glance, at least the videos (posted about six months apart) appear to be VERY similar:

https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-416900


View attachment 70372
View attachment 70373


Now compare the slides to those in the video uploaded October 3:

View attachment 70368


View attachment 70369

View attachment 70370

In fact, when I just tried to cursorily compare the new paper to the March 15 paper that @Fullmoonfever had linked at the time (https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-416313), I discovered that the link he had posted then now connects directly to this new paper, published on September 16, so it seems to be an updated version of the previous paper.

I did notice the addition of another co-author, though: Sebastian Otte, who used to be a PhD student and postdoc at Uni Tübingen (2013-2023) and became Professor at Uni Lübeck’s Institute for Robotics and Cognitive Systems just over a year ago, where he heads the Adaptive AI research group.

To put the results that our competitors’ neuromorphic offerings fared worse in the benchmarking tests alongside Akida somewhat into perspective:
In all fairness, it should be highlighted that Akida’s superiority was at least partly due to the fact that AKD1000 is available as a PCIe Board, whereas SynSense’s DynapCNN was connected to the PC via USB and - as the excerpt Gazzafish already posted shows - the researchers did not have direct access to a Loihi 2 edge device, but merely through a virtual machine provided by Intel via their Neuromorphic Research Cloud. The benchmarking would obviously yield better comparable results if the actual hardware used were of a similar form factor:

“Our results show that the better a neuromorphic edge device is connected to the main compute unit, e.g., as a PCIe card, the better the overall run-time.”


Anyway, Akida undoubtedly impressed the researchers, and as a result they are considering further experiments: “(…) future work could involve evaluating the system with an additional Akida PCIe card.”


View attachment 70374

---

In an earlier post (https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-426404), I had already mentioned that the paper’s first author, Andreas Ziegler, who is doing a PhD in robotics and computer vision at Uni Tübingen, has meanwhile completed his internship at Sony AI in Switzerland (that - as we know - partially funded the paper’s research):

View attachment 70375


Fun fact: One of his co-authors, Karl Vetter, however, is no longer with Uni Tübingen’s Cognitive Systems Lab, but has since moved to France, where he has been working as a research engineer for…

Neurobus for the past three months!
It’s a small world, isn’t it?!

View attachment 70376
View attachment 70377

Three days ago, first author Andreas Ziegler gave a talk on the recent table tennis robot research conducted at Uni Tübingen during the Neuromorphic Vision Hackathon at ZHAW (Zürcher Hochschule für Angewandte Wissenschaften / Zurich University of Applied Sciences), where robotics and neuromorphic computing expert Yulia Sandamirskaya (ex Intel Labs) heads the Research Centre “Cognitive Computing in Life Sciences” at ZHAW’s Wädenswil campus.

View attachment 71872


While the content of his presentation is not new for those of you who already read the paper or saw the video, I thought the way he presented it was quite cool, with all the embedded videos! Have a look yourselves:

Event-Based Vision for Fast Robot Control

Anyway, more exposure for Akida and those favourable benchmarking results (even though it is unclear how much influence the hardware’s form factor had, see my post above).



Here are some of the presentation slides:

View attachment 71874

View attachment 71875
View attachment 71879


View attachment 71877

View attachment 71880

View attachment 71883

View attachment 71884

In Andreas Ziegler’s updated CV (https://andreasaziegler.github.io/), we can now see who his supervisors were during his internship at Sony AI (that funded this research): Raphaela Kreiser and Nagoya Takahashi:

View attachment 71881


View attachment 71882

The revised paper on Uni Tübingen research involving Akida ( ), which was partially funded by Sony AI, got accepted for ICRA 2025, the International Conference on Robotics & Automation in Atlanta (19-23 May).

The results of #ICRA2025 are out, and I'm happy to share that our work "Detection of Fast-Moving Objects with Neuromorphic Hardware" has been accepted. | Andreas Ziegler

The results of #ICRA2025 are out, and I'm happy to share that our work "Detection of Fast-Moving Objects with Neuromorphic Hardware" has been accepted. Project page: https://lnkd.in/ew-U5-BD Preprint: https://lnkd.in/enJWM2Vt (Data and code will follow) This paper's journey started with Karl...

www.linkedin.com

As mentioned previously, the additional co-author of the revised version, Sebastian Otte, who used to be a PhD student and postdoc at Uni Tübingen (2013-2023), became a professor at Uni Lübeck’s Institute for Robotics and Cognitive Systems in September 2023, where he heads the Adaptive AI research group. He also leads the University of Lübeck research team that is collaborating with Mercedes-Benz, Intel and other partners on the NAOMI4Radar project (using Loihi 2). Otte is a uni researcher who has got first-hand experience with both Akida and Loihi, and has also expressed his appreciation for both:

https://thestockexchange.com.au/threads/brn-discussion-ongoing.1/post-438969

The other co-author named in Andreas Ziegler’s LinkedIn post, whose Master’s thesis was the initial basis for the research at Uni Tübingen that was to follow and ultimately led to the paper now accepted for ICRA 2025, is Karl Vetter, who left academia last July to work for BrainChip’s partner Neurobus (who are also partnered with Prophesee and Intel).


After winning first place at the European Defense Tech Hackathon in Paris two months ago…

Gregor Lenz, Florian Corgnou and Karl Vetter from BrainChip’s partner Neurobus were part of a team that came in first at the European Defense Tech Hackathon, which took place in Paris over the weekend.
Their winning solution titled Automatic event-based detection and tracking of UAVs and Shahed drones in challenging lighting conditions “showcased the ground-breaking potential of neuromorphic event-based cameras (…) paving the way for smarter, faster and more efficient defense-systems”.

As you may have guessed from the mentioning of the Iranian-designed Shahed drones (which are also known by their Russian designation Geran-2), the 34 projects in total were far from being destined for storage in an ivory tower of academia: European defense company Helsing AI was a key partner of that hackathon, which was also supported by the Ministry of Defence of Ukraine.

“The challenges were based on real-world problems gathered from our partners, who have delivered solutions to the frontline, from building underwater reconnaissance systems to the interception of Shahed drones and helicopters and swarm coordination in GPS-denied environments.”

HESA Shahed 136 - Wikipedia

en.m.wikipedia.org

View attachment 73720


View attachment 73721

Brave1 - Ukrainian Defense Innovations

brave1.gov.ua

View attachment 73723

… the team from Neurobus must be keen on defending their title at the upcoming European Defense Tech Hackathon in Munich, held in conjunction with the renowned Munich Security Conference:

European Defense Tech Hub

We foster defense innovation in Europe by connecting builders, investors, and policymakers

eurodefense.tech

Hopefully with a little help from Akida?

 

[charles2]

Inferencing....thought to be BRN's strong suit...next frontier of AI

BZAI ....more proactive with news releases

BWI Investment in Blaize (BZAI) Highlights Efficiency as the Key to DeepSeek Advancements
Business Wire
Tue, January 28, 2025 at 6:00 AM CST 2 min read

In This Article:​

BZAI
+8.93%
BZAIW
+13.99%

• With DeepSeek driving market conversations, Blaize’s focus on low-cost Edge AI chips and scalable small language models positions it as a leader in inferencing—the next frontier of AI.
• Trump Calls DeepSeek AI a ‘Wake-Up Call,’ Highlights Need for Innovation in AI Methods

Emphasis mine

 

[DingoBorat]

Nitpicking? Hang on, you literally wrote


hence my comment about the research partnerships they recently announced, and besides, research partnerships can of course eventually lead to the real deal. Why do you think they are doing such research in the first place?! Just for fun? Would they pick partners they could never ever imagine signing a deal with? And how about Akida in the EQXX - was that not research that could potentially lead to the signing of an IP license?

At least we can agree on “So we continue to speculate.”

We’ll likely need quite a bit of patience, though, as the journalists who were recently shown around MB’s Future Technologies Lab, were told by MB’s neuromorphic researchers that the technology were “still in its infancy” and required “extensive testing and certification” before going into production cars, saying they were expecting the hardware to be available in the 2030s…

cf references to various online articles





Which also means that BrainChip’s current advantage of already having a commercially available neuromorphic chip or IP is no longer going to be that relevant in a few years’ time. For the time being, research chips will do…

"We’ll likely need quite a bit of patience, though, as the journalists who were recently shown around MB’s Future Technologies Lab, were told by MB’s neuromorphic researchers that the technology were “still in its infancy” and required “extensive testing and certification” before going into production cars, saying they were expecting the hardware to be available in the 2030s…"

Although I'm not really excited, about what I consider to be a dwindling industry (Great publicity, but) I don't think their "expecting the hardware to be available in the 2030s…" statement is a blanket one, relating to all use cases of Neuromorphic Technology, in their cars.

I think it's more likely, that that particular statement, refers to ADAS type applications.

It just doesn't make sense to me, to assume that it relates to things like voice control and infotainment type applications also, because they don't require "certification" for a start..

 

[Pom down under]

Having just read through the Sean post over at the crapper I’ve come to the conclusion that if anyone wants to block 90% of the posters that constantly find a way off attacking the company, then nows your chance.

 

[Meatloaf]

"We’ll likely need quite a bit of patience, though, as the journalists who were recently shown around MB’s Future Technologies Lab, were told by MB’s neuromorphic researchers that the technology were “still in its infancy” and required “extensive testing and certification” before going into production cars, saying they were expecting the hardware to be available in the 2030s…"

Although I'm not really excited, about what I consider to be a dwindling industry (Great publicity, but) I don't think their "expecting the hardware to be available in the 2030s…" statement is a blanket one, relating to all use cases of Neuromorphic Technology, in their cars.

I think it's more likely, that that particular statement, refers to ADAS type applications.

It just doesn't make sense to me, to assume that it relates to things like voice control and infotainment type applications also, because they don't require "certification" for a start..

I agree with you. First of all, you are correct in stating that the car manufacturing is a dwindling industry. Especially now, considering the costs involved in producing vehicles.
MB stating that neuromorphic is at the infancy stage tells me, that they are working on technology that will be a game changer in the industry.
I don’t consider EV’s to be that game changer. There are currently arguments whether EV’s are good for the environment due to the fact that to produce the batteries is not very environmentally friendly.

So what could MB be working on besides EV’s with a luxury experience. To me it has to be driverless vehicles. Although, several companies are working on driverless vehicles, it has not been perfected. Google, Tesla and other companies have been working on driverless vehicles for years. What else would require extensive testing and certification?

Currently, I’m not hoping for an IP with a vehicle manufacturer. I believe it would be more beneficial for Brn to score an IP contract with a telecommunications company or a company involved with assistive aids such as: hearing aids, eye wear or technologies that get rid of eye wear, robotics, home appliances etc.

Brn are positioning themselves to play a part in every industry. Partnerships, although time consuming, are what will ensure our success in the near future, hopefully. As I have learnt, a lot of patience is required.

The AI industry is developing at such a fast pace atm, it is both scary and exciting.
Who was it that said:
“We live in exciting times”.

Just my opinion

 

[yogi]

Welcome to the Keyword Spotting Revolution

As smart devices, wearables, and consumer appliances increasingly rely on voice interaction, they face a critical challenge: efficiently managing energy consumption to support these interactions.

Keyword spotting technology offers a game-changing solution by enabling devices to process data and respond only after a specific keyword or phrase is detected. This not only extends battery life but also enhances privacy by minimizing unnecessary data processing.

Discover how BrainChip’s innovative technology is transforming industries -from consumer wearables and healthcare to smart home appliances and beyond.

LeadGen KWS - LP

Welcome to the Keyword Spotting Revolution: Where Intelligent Voice Interaction meets Real-World Applications Smart devices, wearables and even consumer appliances are becoming increasingly dependant on voice interaction, yet technology solutions are challenged with efficiently controlling the...

bit.ly

Welcome to the Keyword Spotting Revolution | BrainChip

Welcome to the Keyword Spotting Revolution As smart devices, wearables, and consumer appliances increasingly rely on voice interaction, they face a critical challenge: efficiently managing energy consumption to support these interactions. Keyword spotting technology offers a game-changing...

www.linkedin.com

 

[manny100]

OpenAI warns DeepSeek 'distilled' its AI models, reports

OpenAI has accused Chinese AI start-up DeepSeek of using its proprietary models to train an open-source competitor.

www.newsweek.com

Details of reports DeepSeek is suspected of 'distillation' which is used to smaller models to kebmverage offlarger more advanced models.
Well worth a read.

 

[Sosimple]

Linkin

 

[Tothemoon24]

Linkin

Welcome to the Keyword Spotting Revolution: Where Intelligent Voice Interaction meets Real-World
Applications​

Smart devices, wearables and even consumer appliances are becoming increasingly dependant on voice interaction, yet technology solutions are challenged with efficiently controlling the energy required to support these very specific interactions within the devices.

How to Determine when Device
interaction is Needed ​

Identify Keywords to match​

Are they recognized?​

Device interacts and listens.​

If the keywords are not recognized, the device does not attempt to process any data, which elongates battery life and provides an additional layer of user privacy.

This is Keyword Spotting (KWS)​

Devices detect specific wake words and phrases and interact only when there is a match.

• Reduces User Latency
• Ensures Privacy
• Improves Power Efficiency
• Significantly Reduces

This technology powers the functionality behind familiar voice-activated systems, making them responsive and user-friendly. However, as devices evolve and user expectations grow, Keyword Spotting is now occurring locally on devices, without cloud connectivity.
You no longer need to be a large cloud provider to offer a “hello” and wake up service for devices.

Why Keyword Spotting?​

Keyword Spotting is a technology that listens for specific words or phrases in an audio stream, triggering a response from a device. Once known only for waking up a device service that routes the rest of the voice from your home or smartphone device to a server in the cloud, Keyword Spotting can be deployed locally to interpret a wide set of in context commands, such as “lock the door” or “set the microwave for 1 minute.”
The key is the device’s ability to locally monitor audio input in a low-power state until it detects the specified keyword, allowing the system to save battery or compute power until needed, which is particularly important for devices such as wearables and other IoT devices. Having a small library of key words enables local control of a device without any other interface.

Key Benefits​

Lower Cost of Ownership​

Local edge processing allows devices to operate without relying on cloud connectivity, reducing operational costs.

Reduced Latency​

Responses are nearly instantaneous, providing a natural voice interface, without the need for round trips to the cloud and back.

Enhanced Privacy​

Audio data stays on the device, there’s no risk of personal information being eavesdropped on by remote servers.
There are a host of examples of how this can benefit users in scenarios like a vehicle,where commands like “turn on the lights” or “check for maintenance” offer a betterexperience than searching for switches or menus. It enables instant, connection-freeresponses, which is crucial for safety.

 

[charles2]

Devices detect specific wake words and phrases and interact only when there is a match.

• Reduces User Latency
• Ensures Privacy
• Improves Power Efficiency
• Significantly Reduces???????

Significantly reduces WHAT?

A version of Keyword Spotting might edit errors like this.