BRN Discussion Ongoing

[White Horse]

Anil Mankar likes this





Raul Vergara • 3rd+ Executive Vice President Growth at Edge Impulse
5h • Edited •
5 hours ago

I am very excited with the response for this year's Edge ML series. As we get closer to the event I wanted to share with you a preview of the agenda (check the link!) great content from Demystifying Edge ML to hands on workshops using the best HW out there to run your IoT + ML Applications (TI, Nordic, Alif and brainchip). If you are not technical but still want to learn what it takes to build, join me and one of out experts and let's walk together through the workshops and get an inside view.

March 27 is just around the corner for San Jose and Amsterdam! register int he link below (now with agenda!!!!)
https://lnkd.in/g5aNn6rn
#ml #iot #edge


The program on the Edge Impulse site also mentions Sony.

12:00–13:00 Lunch (demo area)‍
Demos from TI, BrainChip, Alif, Nordic, Sony, MemryX, and NovTech

 

[MDhere]

I was reading something interesting today, well it was interesting to me. It is said that Toyota have a philosophy of asking the why question 5 times on their manufacturing problem solving.
I'm assuming they found their answers now.

 

[Fact Finder]

Anil Mankar likes this





Raul Vergara • 3rd+ Executive Vice President Growth at Edge Impulse
5h • Edited •
5 hours ago

I am very excited with the response for this year's Edge ML series. As we get closer to the event I wanted to share with you a preview of the agenda (check the link!) great content from Demystifying Edge ML to hands on workshops using the best HW out there to run your IoT + ML Applications (TI, Nordic, Alif and brainchip). If you are not technical but still want to learn what it takes to build, join me and one of out experts and let's walk together through the workshops and get an inside view.

March 27 is just around the corner for San Jose and Amsterdam! register int he link below (now with agenda!!!!)
https://lnkd.in/g5aNn6rn
#ml #iot #edge


The program on the Edge Impulse site also mentions Sony.

12:00–13:00 Lunch (demo area)‍
Demos from TI, BrainChip, Alif, Nordic, Sony, MemryX, and NovTech

Sixty thousand ENGINEERS being reminded by Edge Impulse that Brainchip exists.

That one percent is looking achievable.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Steve10]

In 2021, about 1.15 trillion semiconductor units were shipped all over the world - another year where shipments exceeded one trillion. Semiconductors are a crucial element of modern electronic devices. In 2021, the global semiconductor industry registered a year-on-year revenue growth rate of 26.2 percent, substantial growth following the sharp decline in 2019. Growth is expected again in 2022, with forecasts suggesting a figure of around 10.4 percent.

Samsung, the world’s largest semiconductor chip maker as of 2021, had 12.3 percent of the total semiconductor market share. The company's biggest competitor in the market is Intel who occupied 12.2 percent of the market in 2021. Intel’s main business lies in the design and manufacturing of semiconductor chips, accounting for a significant portion of Intel's net revenue.

Global semiconductor unit shipments 2021 | Statista

In 2021, about 1.15 trillion semiconductor units were shipped all over the world - another year where shipments exceeded one trillion.

www.statista.com

1% of 1.15 trillion is a lot of chips.

ARM sold 8 billion last quarter x 4 = 32B / 1.15T = 2.78%.

Samsung's 12.3% market share = 141.45B chips & Intel's 12.2% market share = 140.3B chips.

Samsung's annual revenue is AUD $344,579,703,720 & the MC is AUD $462,103,694,917.

Intel's annual revenue is AUD $93,146,917,500 & the MC is AUD $161,356,167,000. Intel may have smaller market share now as revenue was AUD $116,740,197,000 in 2021.

1,150,000,000,000 chips per year / 365 days = 3,150,684,931 chips per day / 24 hours = 131,278,538 per hour / 60 = 2,187,975 per minute / 60 = 36,466 chips being shipped every second globally.

 

[White Horse]

Nice accolade.
Love the press.




Rob Telson likes this





PROPHESEE 7,446 followers
1d •
1 day ago



We won Best of #MWC23 !
Our collaboration with Qualcomm Snapdragon to unlock a new image quality paradigm for #smartphone photography was the talk of #MWC this year including winning #BestofMWC2023 from Android Authority.

MWC23 was a blast, thank you for this warm welcome and stay tuned for more!

Read all about it
Reuters https://lnkd.in/eCytgKqw
VentureBeat https://lnkd.in/eg7rFFAj
Android Authority https://lnkd.in/e5DXcrUQ
https://lnkd.in/eC2cJzRq
AndroidPolice.com https://lnkd.in/ePpUpFKu
Fierce Electronics https://lnkd.in/eFJwhADG
Le Figaro https://lnkd.in/e-HnddFz
L'Usine Digitale https://lnkd.in/e5Aw6p3u
Imaging and Machine Vision Europe https://lnkd.in/eGErd_ny
New Electronics https://lnkd.in/ePFXytyF
eeNews Europe https://lnkd.in/eD8VB63d

And more here https://lnkd.in/ekgwyCUg

#eventcamera #eventbasedvision #mobiletechnology #mobilephone #photography

 

[Quatrojos]

Our friends at Carnegie Melon are hard at work:

Can we build Cortical Columns Computing Systems (C3S) that possess brain-like capabilities as well as brain-like efficiency?

This chapter presents ongoing research in the Neuromorphic Computer Architecture Lab (NCAL) at Carnegie Mellon University (CMU) focusing on addressing this question. Our initial findings indicate that designing truly intelligent and extremely energy-efficient C3S-based sensory processing units, using off-the-shelf digital CMOS technology and tools, is quite feasible and very promising, and certainly warrants further research exploration.

Cortical Columns Computing Systems: Microarchitecture Model, Functional Building Blocks, and Design Tools

Reverse-engineering the human brain has been a grand challenge for researchers in machine learning, experimental neuroscience, and computer architecture. Current deep neural networks (DNNs), motivated by the same challenge, have achieved remarkable results in Machine Learning applications...

www.intechopen.com

STDP seems to be the best way forward. We have our fingers firmly around the AI bottleneck...

 

[Quatrojos]

This new trillion-dollar sector of the economy will be based onseveral key technologies which are coming to fruition at the same time: A) algorithmicdevelopments in artificial intelligence and machine learning, B) AI chips which are afundamental change in the nature of computing hardware, shifting from von Neumannarchitecture to distributed memory and computation mimicking neural structures of the brain (seefor example research and commercial projects in neuromorphic computing, Brainchips Akida,Intels Loihi, or Graphcore), C) 3D Sensors lidars, radars, and ultrasonic sensors can produce3D point clouds (azimuthal, elevation, and range), providing a new basis for precisionpositioning, object detection, classification, and tracking, and D) the electrification andoptimized coordination of vehicle fleets

(NEPTUNE) Threat Detection & Tracking for Vehicle Gunners - Dimensions

ABSTRACT: Vehicle gunners need to improve the speed and reliability of threat detection and targetacquisition/engagement. Lidar specifically designed for A) long-range, B) rapid, accurate, andflexible beam-steering, C) detection using laser pulse shapes, D) complementing cameraimaging, E)...

app.dimensions.ai

 

[Quatrojos]

Today’s smart devices, robots, and vehicles are becoming ever more autonomous and this places utmost importance on their reliability and safety. To meet these safety demands, light detection and ranging (LiDAR) systems have been used to make a 3D map of environment in order to navigate the autonomous agent and avoid collisions. A LiDAR measures the distance by illuminating a target with laser light and detects the reflection with a sensor. They are becoming an inevitable part of autonomous vehicles, drones, and robots by providing this vital sensing and imaging capability. However, today’s LiDARs also impose potential human and public safety threats due to their security vulnerabilities. For instance, an attacker can deliberately send a spoofing signal to the victim’s LiDAR which cannot differentiate the spoofing signal from the actual reflected signal. In doing so, attacker can overwrite the actual reflected signal. Eventually, the attacker can trick the victim by hiding or misrepresenting its actual location, leading to serious security and safety issues. While LiDAR systems are on the verge of commercialization, these scenarios are unavoidable and the prevention techniques have not been well studied and researched. This project aims at investigating these issues and proposing a new secure scheme based on frequency encryption.

We have a sweet spot with LIDAR...

EAGER: SARE: Secure LiDAR Systems with Frequency Encryption - Dimensions

Today’s smart devices, robots, and vehicles are becoming ever more autonomous and this places utmost importance on their reliability and safety. To meet these safety demands, light detection and ranging (LiDAR) systems have been used to make a 3D map of environment in order to navigate the...

app.dimensions.ai

 

[White Horse]

New article about 1500 tape-out.
https://www.electronics-lab.com/bra...-chip-on-globalfoundries-22nm-fd-soi-process/

 

[Quatrojos]

Machine learning (ML) algorithms have been applied successfully for many years to databases of visual imagery for the recognition of objects in a scene. Sensors working at radio frequency (RF) mostly produce low-resolution data where the signals detected are more abstract and require algorithmic processing to present the information to the human operator. RF sensors can operate actively (detecting reflected signals from objects illuminated by their own transmission) and passively (intercepting emissions and reflections from objects). The characteristics of a target can depend on both reflections and interceptions of RF emissions, which are likely to be at different frequencies and are often context dependent. Long-established classical detection methods tend to work by removing everything that does not look like the signal of interest, which may throw away valuable information and context in the process. This process is typically performed on each type of data separately and only the processed outputs are combined. Real-time co-processing of multiple sensor data streams is analytically and computationally challenging. Fusion of this data usually does not occur until after the individual classical detection processes. However, ML techniques may be able to learn to extract the beneficial features from sensor data of the scene to detect all objects of interest and provide classification of object types earlier in the processing than is usually possible, and with improved confidence compared to classical methods. This PhD will seek to establish whether multi-modal airborne sensor data, used early in the signal processing chain, can be used to detect and classify objects earlier in the processing chain leading to improved detection and classification performance.

Machine Learning for Target Detection and Classification Using Multi-Modal Airborne Sensor Data - Dimensions

Machine learning (ML) algorithms have been applied successfully for many years to databases of visual imagery for the recognition of objects in a scene. Sensors working at radio frequency (RF) mostly produce low-resolution data where the signals detected are more abstract and require algorithmic...

app.dimensions.ai

There's been an explosion of research since AKIDA was born...

 

[Diogenese]

New article about 1500 tape-out.
https://www.electronics-lab.com/bra...-chip-on-globalfoundries-22nm-fd-soi-process/

Great find White Horse,

EDIT: A note of caution - this is not an official BrainChip release, so variations from the 29 january press release may just be journalistic embellishments.

This is new:

BRAINCHIP ANNOUNCES SUCCESSFULLY TAPED OUT AKD1500 CHIP ON GLOBALFOUNDRIES’ 22NM FD-SOI PROCESS​

...
The AKD1500 is BrainChip’s flagship product designed to deliver AI processing capabilities on edge. The chip features BrainChip’s patented spiking neural network (SNN) technology, which is capable of learning, recognizing, and processing patterns in real-time. The AKD1500 is ideal for various applications, including advanced driver assistance systems (ADAS), surveillance, and autonomous robotics.

...
The AKD1500 chip is expected to be available soon and will be a significant addition to BrainChip’s portfolio of edge AI solutions. The company has already received strong interest from several customers looking to use AKD1500 in their products.


No mention of flagships or ADAS in the 20230129 announcement:
https://brainchip.com/brainchip-tapes-out-akd1500-chip-in-globalfoundries-22nm-fd-soi-process/ :
The AKD1500 is designed as an accelerator reference chip, using quad/octal SPI interfaces for MCU modules and PCIe for extensible cards, which assists partners in developing and demonstrating their solutions as a stepping stone to integrating the Akida IP into their production SoCs.

Back to the new announcement:

"customers looking to use the AKD 1500 in their products" -
does this mean that the hardware chips will be sold as commercial products rather than only as a demonstration on a PCIe board?

MegaChips can provide the Akida 1000 SoC chip, and GF can supply the AKD 1500 partial SoC for which "The Company has already received strong interest from several customers"?

If there is a market for the Akida 1500 chip, this could prove a faster way to market than the IP licence route: "The AKD 1500 chip is expected to be available soon"

 

[DingoBorat]

Great find White Horse,

This is new:

BRAINCHIP ANNOUNCES SUCCESSFULLY TAPED OUT AKD1500 CHIP ON GLOBALFOUNDRIES’ 22NM FD-SOI PROCESS​

...
The AKD1500 is BrainChip’s flagship product designed to deliver AI processing capabilities on edge. The chip features BrainChip’s patented spiking neural network (SNN) technology, which is capable of learning, recognizing, and processing patterns in real-time. The AKD1500 is ideal for various applications, including advanced driver assistance systems (ADAS), surveillance, and autonomous robotics.

...
The AKD1500 chip is expected to be available soon and will be a significant addition to BrainChip’s portfolio of edge AI solutions. The company has already received strong interest from several customers looking to use AKD1500 in their products.


No mention of flagships or ADAS in the 20230129 announcement:
https://brainchip.com/brainchip-tapes-out-akd1500-chip-in-globalfoundries-22nm-fd-soi-process/ :
The AKD1500 is designed as an accelerator reference chip, using quad/octal SPI interfaces for MCU modules and PCIe for extensible cards, which assists partners in developing and demonstrating their solutions as a stepping stone to integrating the Akida IP into their production SoCs.

Back to the new announcement:

"customers looking to use the AKD 1500 in their products" -
does this mean that the hardware chips will be sold as commercial products rather than only as a demonstration on a PCIe board?

MegaChips can provide the Akida 1000 SoC chip, and GF can supply the AKD 1500 partial SoC for which "The Company has already received strong interest from several customers"?

If there is a market for the Akida 1500 chip, this could prove a faster way to market than the IP licence route: "The AKD 1500 chip is expected to be available soon"

"If there is a market for the Akida 1500 chip, this could prove a faster way to market than the IP licence route: "The AKD 1500 chip is expected to be available soon"

Personally, I don't think the Company has changed strategy and are just testing the "recipe" as Rob put it, but it could be a bit of both?..

These customers would be very sizable and it seems a bit odd, that they would take chips, rather than have more control over the options of how many nodes etc?..

The early timing of the capital raise, does give weight, to the possibility of a production run of AKIDA1500 though..

I'm a bit back and forth on this one ..

I'd prefer they stuck to the IP model..
If the chip run was for customers, wouldn't it be a straight up payment?..

If it's a case of "you put up the funds and run a large batch and if they work, we might buy them" scenario..
Then it seems like a bit of a high risk play..


On another thought..

A new IP licence, that we could probably lay money on, would be MegaChips, as their current licence only covers AKIDA1000.

 

[Diogenese]

"If there is a market for the Akida 1500 chip, this could prove a faster way to market than the IP licence route: "The AKD 1500 chip is expected to be available soon"

Personally, I don't think the Company has changed strategy and are just testing the "recipe" as Rob put it, but it could be a bit of both?..

These customers would be very sizable and it seems a bit odd, that they would take chips, rather than have more control over the options of how many nodes etc?..

The early timing of the capital raise, does give weight, to the possibility of a production run of AKIDA1500 though..

I'm a bit back and forth on this one ..

I'd prefer they stuck to the IP model..
If the chip run was for customers, it would have to be a straight payment?..

If it's a case of "you put up the funds and run a large batch and if they work, we might buy them" scenario..
Then it seems like a bit of a high risk play..


On another thought..

A new IP licence, that we could probably lay money on, would be MegaChips, as their current licence only covers AKIDA1000.

MegaChips do not have the FD-SIO technology.

Edit: per @Tothemoon24 and @DingoBorat , MegaChips were involved in the design of AKD1500.

 

[DingoBorat]

MegaChips do not have the F-SIO technology.

I thought they were more of a design house, like ARM and the actual chips are produced at foundries?

Or that's correct, but they don't have licence to work with the other design process?

 

[DingoBorat]

I thought they were more of a design house, like ARM and the actual chips are produced at foundries?

Or that's correct, but they don't have licence to work with the other design process?

Regardless though, the AKIDA1500 IP would also be process agnostic?

Or I guess we haven't actually proven that, with AKIDA1000 yet?

I'm guessing it has a few more bells and whistles over AKIDA1000 and is not just the IP minus ARM processor and on another chip architecture?

 

[DingoBorat]

Regardless though, the AKIDA1500 IP would also be processor agnostic?

I'm guessing it has a few more bells and whistles over AKIDA1000 and is not just the IP minus ARM processor and on another chip architecture?

Some of my terminology is probably a bit jumbled, but I'm sure you know what I mean

 

[Tothemoon24]

Apologies if this is common knowledge?





Megachips played a part in design of Akida1500

 

[Deleted member 118]

Old video but very interesting to listen to regarding disclosure, especially as it’s been discussed a fair bit here recently

Won’t let me post the link btw so I’ve added a screen shot if anyone can assist

https://***************.com.au/brai...erns-around-continuous-disclosure-2022-02-02/

 

[MDhere]

Good morning fellow brners, now can someone tell me the difference when someone states these words or someone states these words -

Aim to "commercialise"products

or

Aim to " launch" products

which represents (rocket ship ready for the button to be pressed ) which is it

 

[Labsy]

"If there is a market for the Akida 1500 chip, this could prove a faster way to market than the IP licence route: "The AKD 1500 chip is expected to be available soon"

Personally, I don't think the Company has changed strategy and are just testing the "recipe" as Rob put it, but it could be a bit of both?..

These customers would be very sizable and it seems a bit odd, that they would take chips, rather than have more control over the options of how many nodes etc?..

The early timing of the capital raise, does give weight, to the possibility of a production run of AKIDA1500 though..

I'm a bit back and forth on this one ..

I'd prefer they stuck to the IP model..
If the chip run was for customers, wouldn't it be a straight up payment?..

If it's a case of "you put up the funds and run a large batch and if they work, we might buy them" scenario..
Then it seems like a bit of a high risk play..


On another thought..

A new IP licence, that we could probably lay money on, would be MegaChips, as their current licence only covers AKIDA1000.

I suspect these "customers" are keen to test the chips functionality and validate 1500 with their particular needs before signing onto Ip licencing. My guess is government contracts, hence use of GF. Defence! And not just USA. Possibly global defence contractors friendly with USA. Through Megachips. They are keen on flushing US market.
Just a hunch.