BRN Discussion Ongoing

[Jumpchooks]

Swiss cement supplier Holcim to order 1,000 Mercedes-Benz eActros 600 electric trucks - Charged EVs

Swiss building materials producer Holcim has signed a joint letter of intent with Mercedes-Benz Trucks to purchase 1,000 eActros 600 electric trucks, the majority of which will be used with silo trailers for transporting building materials, including cement. It is the largest planned single...

chargedevs.com

 

[SebThatGermanChap]

Use Trading Tools to Find More Trading Opportunities

Use our high-performing trading tools to find the most favourable opportunities in the markets and boost your earning potential.

www.xm.com

 

[IloveLamp]

We get a tag amongst some big names here, and Rob likes it.

Not a small company.......10k employees, she is very excited about neuromorphic computing.

Ronit Gabizon on LinkedIn: #neuromorphiccomputing #ai #futureoftech #nvidia #intel #ibm #microsoft…

The Neuromorphic Revolution: Redefining AI and Problem-Solving! Hey LinkedIn community! Brace yourselves for the neuromorphic revolution, a cutting-edge field…

www.linkedin.com

IEC - Israel Electric Corporation חברת החשמל לישראל בע"מ | LinkedIn

IEC - Israel Electric Corporation חברת החשמל לישראל בע"מ | 20,076 followers on LinkedIn. The Israel Electric Corporation (IEC) is a public, government company, 99.85% of its shares are government-owned and traded on the Tel-Aviv and Singapore stock exchange. Company activities include...

www.linkedin.com

 

[Chris B]

Hi Chris
Not 100% sure but think we may have seen an earlier version of this going by the summary:

BrainChip's neuromorphic technology: The next wave of AI @ CES 2024​

PLUGHITZ Live
3.32K subscribers
Join

8 waiting Premieres Feb 8, 2024 #ces2024 #ceslive #techpodcasts
#ces2024 #ceslive #techpodcasts BrainChip is changing how we think about AI by developing a neural network accelerator that mimics the human brain's ability to multitask. http://plughitzlive.com/ces A transformative shift in the way devices are perceived and interacted with is being brought about by AI neural network accelerators. BrainChip Inc. (https://brainchip.com/) is a leading developer of such an accelerator, a technology that enables AI to be embedded into various devices, thereby facilitating intelligent functions.

AI neural network accelerators: A technological revolution BrainChip's neural network accelerator is compared to other AI technologies in the market, such as those developed by NVIDIA. The distinguishing factor for BrainChip is its architecture, which is based on neuromorphic principles, akin to the human brain. This implies that the accelerator is capable of multitasking and processing multiple events simultaneously, much like the human brain. Furthermore, the accelerator is noted for its minimal energy consumption, contributing to its high efficiency. The end product for consumers is a device that requires less frequent charging, thanks to its neuromorphic architecture. This technology enables devices to perform multiple functions concurrently, similar to the human brain. Take, for example, a smart cabin in a vehicle, where the accelerator can recognize the driver, gestures, and voice. This level of recognition and interaction is made possible by BrainChip's technology. AI still has room to grow However, it is important to note that the adoption of this technology is still underway. The development and refinement of AI technologies require time, as they need to achieve a level of accuracy that can be utilized by consumers on a daily basis. It is the combination of inputting correct data and understanding it that facilitates the development of AI-based applications. Excitement about the future of AI and the demand for more intelligent devices from consumers is a constant. At CES 2024, AI has been a prevalent theme, with numerous companies and individuals focusing on delivering intelligent solutions. Whether it is through software, hardware, or embedded intelligence, AI is transforming various industries and products.

Rob Telson, the Vice President of Ecosystem and Partnerships at BrainChip Inc., acknowledges that AI has only been around for a relatively short period of time, with significant advancements occurring in the past three years. He specifically mentions GPT, or Generative Pre-trained Transformer, which is a language model that has gained attention for its ability to generate human-like text. Despite these advancements, Linsdell emphasizes that truly intelligent devices that can think at the same level as a human are still three to seven years away. Of course, the big question in AI whether AI will ever surpass the human brain in intelligence. Linsdell firmly states that AI will not be able to achieve the same level of thinking as the human brain. He acknowledges that AI may be able to think ahead of humans in certain areas, but it still lacks the efficiency, rationality, and logic that the human brain possesses. However, Linsdell also believes that AI has the potential to bring about significant benefits if used correctly. He mentions the importance of playing the game right and ensuring that AI is used in a way that aligns with human thinking. While AI may make decisions that are not in line with human thinking currently, Linsdell suggests that with further tinkering and advancements, AI may eventually reach a point where it can match human thinking.

Conclusion: BrainChip is on the bleeding edge of AI processing In conclusion, AI neural network accelerators, such as the one developed by BrainChip Inc. (https://brainchip.com/), are revolutionizing the capabilities of devices. By embedding AI into devices, they can perform multiple functions simultaneously, akin to the human brain. This technology holds the potential to enhance various industries and create more intelligent and interactive devices for consumers.

My opinion only DYOR
Fact Finder

Hi FF.. fooled me by saying it was premiering ;-)

 

[JB49]

https://arxiv.org/html/2401.06911v1

"Furthermore, due to the runtime limitation of 20 minutes for jobs on Intel’s cloud, on-chip training was not possible, restricting the exploration of this important capability of Loihi".

Akida can solve that.

 

[Smoothsailing]

Robs comments regarding Brainchip ability to control the drivers speed range due to the age of the driver was very interesting and something all cars should adopt.

 

[Diogenese]

Robs comments regarding Brainchip ability to control the drivers speed range due to the age of the driver was very interesting and something all cars should adopt.

Yes. You should be limited to twice your age.

 

[Taproot]

Arm shares soar as much as 41% after chip designer gives strong forecast, says AI is boosting sales​

Arm shares soar as much as 41% after chip designer gives strong forecast, says AI is boosting sales

Arm's intellectual property is in nearly every smartphone, many PCs, and thousands of miscellaneous other chips.

www.cnbc.com

ARM Market Cap is going to reach 100 Billion USD by the end of the week at this rate.
Outstanding stuff.

BrainChip

Ubiquitous, distributed AI.<br/>Close to the sensor.<br/>Inspired by the human brain. <br/>

www.arm.com

BrainChip Joins Arm AI Partner Program

BrainChip joins the ARM AI Partner Program, bringing together the power of BrainChip's AI solutions with the industry-leading ARM ecosystem.

brainchip.com

BrainChip integrates Akida with Arm Cortex-M85 Processor, Unlocking AI Capabilities for Edge Devices

Explore how this integration revolutionizes the field of artificial intelligence, enabling efficient and powerful AI processing at the edge.

brainchip.com

https://armkeil.blob.core.windows.net/developer/Files/pdf/ai-ecosystem-catalogue/BrainChip_Arm-AI-Partner-Solution-Brief-Template-1.pdf

BrainChip Taps Former ARM Executive Antonio J. Viana as Non-Executive Director

BrainChip Holdings Ltd (ASX: BRN), (OTCQX: BRCHF), a leading provider of ultra-low power, high-performance artificial intelligence technology, today announced the appointment of Antonio J. Viana as a non-executive director.

www.design-reuse.com

BrainChip Names Former Arm Executive Nandan Nayampally as Chief Marketing Officer

Learn about the leadership and expertise that Nandan brings to BrainChip as they continue to revolutionize the field of AI.

brainchip.com

I could go on and on and on , but I think you get the gist.
If BrainChip can latch onto 1% of ARM's action, we're golden !

 

[7für7]

Let me guess… the shorter are back and want to fish every penny they can get from the last bounce? It’s ridiculous! Or classic brainchip manipulation!

 

[buena suerte :-)]

Yes. You should be limited to twice your age.

Dio.. ?? Would 'you' be breaking the speed limit??

 

[Esq.111]

Afternoon Chippers ,

On the three day chart @ 1 , 2 & 3 min time duration on volume, thinking we should see some decent volume within next hour .... i shall refrain from a price guess today .

.

Regards,
Esq

 

[David holland]

Let me guess… the shorter are back and want to fish every penny they can get from the last bounce? It’s ridiculous! Or classic brainchip manipulation!

Watch the Financials

 

[Diogenese]

Dio.. ?? Would 'you' be breaking the speed limit??

Not on an autobahn.

 

[TheFunkMachine]

https://www.linkedin.com/posts/tcs-...3-5nQd?utm_source=share&utm_medium=member_ios

Some progress being made with TATA using neuromorphic technology in cube sats!

Anil Mankar commented on this post saying “very cool”.

There is no reason to doubt the engagement of our announced partners, however it is always nice with a little reminder and progress update.

The benchmark results is amazing as well and works well in our favour as people in the industry should know that Akida is the brains behind these results

 

[buena suerte :-)]

Not on an autobahn.

Good answer...

 

[7für7]

Watch the Financials

I don’t understand Netherlandish i already told you that.

Home | The Financials

www.thefinancials.nl

 

[Fact Finder]

We get a tag amongst some big names here, and Rob likes it.

Not a small company.......10k employees, she is very excited about neuromorphic computing.

Ronit Gabizon on LinkedIn: #neuromorphiccomputing #ai #futureoftech #nvidia #intel #ibm #microsoft…

The Neuromorphic Revolution: Redefining AI and Problem-Solving! Hey LinkedIn community! Brace yourselves for the neuromorphic revolution, a cutting-edge field…

www.linkedin.com

IEC - Israel Electric Corporation חברת החשמל לישראל בע"מ | LinkedIn

IEC - Israel Electric Corporation חברת החשמל לישראל בע"מ | 20,076 followers on LinkedIn. The Israel Electric Corporation (IEC) is a public, government company, 99.85% of its shares are government-owned and traded on the Tel-Aviv and Singapore stock exchange. Company activities include...

www.linkedin.com

View attachment 56241
View attachment 56238 View attachment 56239 View attachment 56240

After reading this post adding in the ARM partnership, the CRO Edge Impulse statement of AKIDA in every semiconductor and Sean Hehir CEO stating Brainchip intends to be one of the two to three major players in this space I think it is safe to conclude one percent of the Edge Ai market is the least Brainchip will achieve.

My opinion only DYOR
Fact Finder

 

[Diogenese]

Good answer...
View attachment 56252

Is that all?

 

[Diogenese]

I don’t understand Netherlandish i already told you that.

Home | The Financials

www.thefinancials.nl

You have to double it.

 

[Fact Finder]

https://arxiv.org/html/2401.06911v1

"Furthermore, due to the runtime limitation of 20 minutes for jobs on Intel’s cloud, on-chip training was not possible, restricting the exploration of this important capability of Loihi".

Akida can solve that.

Hi JB

This is an important find. I have taken the following extract to cover the facts you have revealed:

• •
  Superiority of SNN and Loihi 2: The results highlight that, in all cases, Intel’s Loihi 2 performs better than the CNN implemented in Xilinx’s VCK5000. However, it is worth noting that, as the batch size increases, the advantage of Loihi became less pronounced as the performance points move closer to the EDP line, although it still outperforms the Xilinx’s VCK5000 implementation.
• •
  Interference detection as a promising use case: Among the considered scenarios, it seems that the interference detection and classification benefited the most from the implementation on Intel’s Loihi chipset. Even though the time ratio remained generally higher than one, the energy savings achieved with Loihi were significant. In some cases, the energy ratio reached values as high as 105
• •
  SNN encoding: Fig. 3 compares the impact of FFT Rate and FFT TEM encoding for the ID scenario. Interestingly, the type of coding used did not have a significant impact on this comparison.
• •
  RRM’s Performance: While RRM did not achieve energy savings as pronounced as in the ID scenario, it consistently achieved an energy ratio exceeding 102

Regarding the digital beamforming performance for fast-moving users, we compared the conventional LASSO solution provided by CVX [13] running on Matlab with the solution of S-LCA on Intel’s Lava simulator. Firstly, it is worth highlighting that the proposed beamforming formulation yielded sparse beamforming vectors, with both solutions being able to turn off up to 60% of the RF chains without compromising the resulting beampatterns. Regarding performance comparisons between the two solutions, both generated satisfactory beampatterns with the main lobe pointing toward the aircraft. For a numerical comparison, the beamformer’s average output power was considered as key performance indicator to assess the beamforming capabilities to mitigate the effects of noise and interference while focusing on the desired signal direction. In this context, the S-LCA solution was able to reach lower levels of beamformer’s average output power, around 19% below the value reached by the CVX solution, but with a much higher spreading of values, around 4 times higher than the CVX solution, when comparing the lower and upper quartiles of beamformer’s average output power.

V-ARemarks about the results obtained with Intel’s Loihi 2​

The results presented in this article were conducted with Loihi 2 but using the remote access from Intel’s research cloud. Although Intel offers the possibility of shipping physical chips to INRC partners premises, at the moment of developing these results the primary access to Loihi 2 was through the Intel’s neuromorphic research cloud. Obviously, the remote access introduced some additional limitations as it was not possible to control for concurrent usage by other users, which could lead to delays and increased power consumption. Additionally, the specific interface used by Intel on their cloud was not disclosed, potentially resulting in differences when conducting measurements with future releases of Loihi 2. Furthermore, due to the runtime limitation of 20 minutes for jobs on Intel’s cloud, on-chip training was not possible, restricting the exploration of this important capability of Loihi.
The cloud interface plays a key role, as it impacts the transfer of spiking signals to the chipset. High input size may span long per-step processing time. For example, in the flexible payload use case, the execution time per example increased from approximately 5 ms to 100 ms when the input size went from 252 neurons to 299 neurons.

VI Conclusion​

While we enter the era of AI, it becomes evident that energy consumption is a limiting factor when training and implementing neural networks with significant number of neurons. This issue becomes particularly relevant for nonterrestrial communication devices, such as satellite payloads, which are in need of more efficient hardware components in order to benefit from the potential of AI techniques.
Neuromorphic processors, such as Intel’s Loihi 2, have shown to be more efficient when processing individual data samples and are, therefore, a better fit for use cases where real world data arrives to the chip and it needs to be processed right away. In this article, we verified this hypothesis using real standard processor solutions, such as the Xilinx’s VCK5000 and the Intel’s Loihi 2 chipset.

Acknowledgments​

This work has been supported by the European Space Agency (ESA) funded under Contract No. 4000137378/22/UK/ND - The Application of Neuromorphic Processors to Satcom Applications. Please note that the views of the authors of this paper do not necessarily reflect the views of ESA. Furthermore, this work was partially supported by the Luxembourg National Research Fund (FNR) under the project SmartSpace (C21/IS/16193290).”

The first Fact is that it confirms that Loihi 1 & 2 are most definitely still only research platforms.

The second Fact is that the lack of willingness of Intel to disclose the Cloud interphase it uses would as the paper suggests be a significant limitation for commercial adoption.

The third Fact disclosed that inference time increased on Loihi 2 the greater the number of neurons in use means is a huge issue for Intel as from every thing I have read about AKIDA the more nodes hence neurons in play the faster the inference times. (Perhaps I miss the point here so this needs a Diogenese ruler run over my conclusion.)

Fact four even Loihi 2 despite its limitations is far more efficient than Xilinx’s VCK5000 running CNN’s.

Finally Fact five using neuromorphic computing for cognitive communication is clearly the way forward.

My opinion only DYOR
Fact Finder