BRN Discussion Ongoing

[hotty4040]

Solid post FF. Glad to see you are back.

Specsavers, Mr Pope, could come in handy, I wish FF had/has returned, or did I miss something, maybe. Have another red, popey, and start praying for his return, imho. I hope all's well with FF, because we miss him heaps.

Anyway, back to the tiges/cats game which miraculously is favoring the tiges atm.

Good week chippers, so please may it continue next week. Some nice reveals over the last few days. Everybody " hang-in-there "

Akida Ballista is gaining momentum gradually.

hotty...

 

[The Pope]

Specsavers, Mr Pope, could come in handy, I wish FF had/has returned, or did I miss something, maybe. Have another red, popey, and start praying for his return, imho. I hope all's well with FF, because we miss him heaps.

Anyway, back to the tiges/cats game which miraculously is favoring the tiges atm.

Good week chippers, so please may it continue next week. Some nice reveals over the last few days. Everybody " hang-in-there "

Akida Ballista is gaining momentum gradually.

hotty...

Hi Hotty,
Yeah maybe you have missed something. It goes back a while when FF came back from a break and stated he had been giving information to others to post he trusted in TSE land and my recall he even quote a couple of names and even said I bet others didn’t notice. Well I did and others did as well. All good either way and FF thoughtful posts are amongst us from time to time. Yes it would be great to have FF officially back but happy with his thoughts being posted by others.

Not having a red tonight, just a few Makers Mark No. 46 on ice. Hope everyone has a great weekend.

Cheers
The pope

 

[Diogenese]

Hi Hotty,
Yeah maybe you have missed something. It goes back a while when FF came back from a break and stated he had been giving information to others to post he trusted in TSE land and my recall he even quote a couple of names and even said I bet others didn’t notice. Well I did and others did as well. All good either way and FF thoughtful posts are amongst us from time to time. Yes it would be great to have FF officially back but happy with his thoughts being posted by others.

Not having a red tonight, just a few markers mark 46 on ice. Hope everyone has a great weekend.

Cheers
The pope

So you haven't got that infallibility thing back from the cleaners yet?

 

[The Pope]

So you haven't got that infallibility thing back from the cleaners yet?

Not yet but when ready to pick up do you want me to pick up yours as well?

 

[Frangipani]

This was posted by Edge Impulse recently

View attachment 12379


STMicroelectronics

STMicroelectronics’ new inertial modules enable AI training inside the sensor - ST News

STMicroelectronics (NYSE: STM), a global semiconductor leader serving customers across the spectrum of electronics applications, has introduced new...

newsroom.st.com

The ISM330ISN always-on 6-axis inertial measurement unit (IMU) for movement and position sensing uses its embedded intelligence to deliver unrivaled performance and accuracy for its size and power. Ideal for IoT and industrial applications, ST’s new IMU sets to accelerate response time and extend battery life in equipment such as condition monitors for predictive maintenance, as well as battery-operated asset trackers and industrial applications such as robots.

The intelligence built into the ISM330ISN enables smart devices to perform advanced motion-detection functions in the sensor without interaction with the external microcontroller (MCU), thus saving power at system level. ST’s approach integrates a specialized processor, the ISPU, in a small area directly on the sensor chip, optimized for machine-learning applications. This enables the ISM330ISN module to have a 50% smaller footprint and consume 50% less power than a typical co-packaged MCU.

ISM330IS - STMicroelectronics

ISM330IS - iNEMO inertial module: always-on 3D accelerometer and 3D gyroscope with embedded ISPU - intelligent sensor processing unit, ISM330ISNTR, ISM330ISTR, STMicroelectronics

www.st.com

View attachment 12377 View attachment 12378

I was going down a rabbit hole after reading about Germany’s third largest automotive supplier ZF Friedrichshafen signing a multi-year contract with STMicroelectronics to purchase SiC modules, in addition to ZF’s partnership agreement for SiC technology with Wolfspeed ( > connection to Brainchip via Duy-Loan Le who is a board member in both companies and by the way also an admirable philanthropist, just like Peter van der Made) announced in February.
Down that rabbit hole, I kept turning my head to check wether I was followed by Diogenese’s infamous ogre whom I imagine to have become a secluded cave hermit (or possibly underground dweller?!) in his retirement, but here is an English article about the recent deal anyway:

ZF signs long-term supply contract for SiC devices with STMicroelectronics

SiC modules from ST will go into ZF's inverter platform from 2025 onwards

www.eenewseurope.com

In the darkness of that underground tunnel system, I suddenly tripped over an upcoming webinar by STMicroelectronics. Any chance Akida could be the reason for the sensors’ intelligence? (My forum search came up with the above post by @M_C last July.) If so, maybe one of our resident tech experts may want to register for one of the sessions and possibly ask clever questions? IST stands for Indian Standard Time (which is UTC + 5:30h), while SGT stands for Singapore Time and is thus the same time zone as Perth, which means the webinar will take place at 5 pm Sydney/Melbourne/Brisbane time on Monday. Alternative Tuesday sessions are offered for participants in Europe/Africa/the Americas.

Webinar: In-sensor monitoring with intelligent MEMS sensors – STMicroelectronics

Discover how you can easily implement accurate in-sensor inclinometer and vibration monitoring applications with an IMU featuring an embedded ISPU.

content.st.com

Live webinar​

In-sensor monitoring with​

intelligent MEMS sensors​

Monday, May 15, 2023​

► 11:30 am IST | 2:00 pm SGT | APAC session​

Tuesday, May 16, 2023​

► 11:00 am CEST | EMEA session​

► 02:00 pm EDT | Americas session​

​

Register now!​

Select the session you are interested in​

Monday, May 15, 2023

► 11:30 am IST | 2:00 pm SGT | APAC session

Register

Tuesday, May 16, 2023

► 11:00 am CEST | EMEA session	Register
► 02:00 pm EDT | Americas session	Register

Sensors with an intelligent sensor processing unit (ISPU) can run calibration and monitoring algorithms without microcontroller intervention. This represents a major step forward in the battery-operated systems in wide scale predictive maintenance scenarios for industrial machinery, or damage and defect detection along large structures like bridges.
In this webinar, we will explore how you can easily implement accurate in-sensor inclinometer and vibration monitoring applications with an IMU featuring an embedded ISPU. This intelligent sensor can implement self-calibration and run sensor fusion algorithms and the sliding discrete Fourier transform (SDFT) for continuous and accurate monitoring on a power budget of a few microwatts.
We will demonstrate the ISPU toolchain at your disposal and the X-CUBE-ISPU libraries to run self-calibration, sensor fusion, and SDFT algorithms immediately on the ISPU.

We will cover:​

• understanding the intelligent sensor processing unit inside a sensor
• the ISPU algorithms available for inclinometer and vibration monitoring
• how to build a vibration monitoring use case using the ISM330IS IMU
• a testimonial of real use-case involving severe vibration

Agenda​

• Industrial applications where smart sensors make the difference
• How one IMU can solve many technical challenges
• The ISM330IS IMU and its edge processing features
• Loading ready-to-use algorithms with Unicleo-GUI and STM32CubeIDE
• A customer testimonial experience with an ISPU sensor
• Conclusions and takeaways

Speaker​

1. 
   
2. 
   

Lorenzo Bracco​

Senior engineer​

Lorenzo is a senior engineer at ST engaged in developing innovative algorithms and tools to support MEMS sensor applications. He also provides key customer support and has filed several patents covering technologies for industrial and personal electronics applications.
PreviousNext

Product details​

• ISM330IS - industrial IMU MEMS sensor with ISPU
• STEVAL-MKI230KA - DIL24 adapter
• X-CUBE-ISPU - software package with ready-to-go libraries
• NUCLEO-F401RE - Nucleo development board
• X-NUCLEO-IKS02A1 - Industrial sensor shield for Nucleo board
• STM32CubeIDE - compiler

———————————————————————————————————-

Here is a press release by STMicroelectronics published yesterday:

https://www.eetasia.com/stmicroelectronics-toolchain-and-software-package-ease-development-of-edge-processing-with-ispus/

STMicroelectronics Toolchain and Software Package Ease Development of Edge Processing with ISPUs​

Article By : STMicroelectronics​

• Category : New Products
• 
  
• 2023-05-11
• 
  
• (0) Comments
• 
  

STMicroelectronics has released a toolchain and accompanying software package for programming the ISPU embedded in the latest-generation intelligent MEMS IMUs.
STMicroelectronics has released a toolchain and accompanying software package for programming the intelligent sensor processing unit (ISPU) embedded in the latest-generation intelligent MEMS IMUs, ISM330IS and LSM6DSO16IS. The toolchain and software help employ the ISPU to handle motion-related workloads such as activity recognition and anomaly detection directly in the sensor. This permits reducing system power and latency, offloading the local microcontroller, and uniquely specializing the behavior of the sensor to the application.

Using the ISPU toolchain, developers can program the sensor’s intelligent processing unit using the familiar and widespread C programming language. They can choose to work from a command line interface (CLI) or an Eclipse-based environment like STM32CubeIDE, and to use a graphical user interface (GUI) such as AlgoBuilder and Unicleo.

The X-CUBE-ISPU software package contains templates and example projects as well as ready-to-use libraries that help developers quickly understand how to use and program the sensors’ ISPU and can be used as a starting point to implement custom algorithms. Pre-built files are also available, which lets users load the X-CUBE-ISPU examples directly into the sensor using one of the GUIs, with no coding required. In addition, a GitHub repository is available providing more examples, tutorials, and other development resources.

Using these resources helps make short work of developing applications such as personal electronics including wearable devices for activity recognition and health monitoring, as well as industrial devices such as asset trackers, equipment-condition monitors, robots, and machine controllers.

ST’s ISM330IS and LSM6DSO16IS inertial modules contain an always-on 3D accelerometer and 3D gyroscope with the embedded ISPU. They feature low power consumption, drawing as little as 0.46mA in low-power mode, and low noise at 70μg/√Hz in high-performance mode. Sensor hub functionality allows them to collect data from up to four additional external sensors. An embedded temperature sensor is also included, and each device is housed in a compact 2.5-by-3-by-0.83mm plastic land grid array (LGA) package.

SPM Instrument, a Strängnäs, Sweden, based innovator in condition monitoring and process optimization, relied on ST’s ISPU toolchain and the X-CUBE-ISPU to tailor, in a short time, the ISPU behavior in their new products for vibration-severity analysis that contain the ISM330IS sensor.

SPM’s sensing solution is ideal for remote condition monitoring of standard production equipment such as pumps and fans, as well as inaccessible machines and equipment placed in hostile or risky environments. The ISM330IS enabled designers to meet a tight power budget, also overcoming the limited processing capabilities available in the local microcontroller.

SPM also features in ST’s intelligent-sensing webinar, which is available at the registration page: Webinar: In-sensor monitoring with intelligent MEMS sensors – STMicroelectronics.

 

[wilzy123]

Yup sure was lol

 

[IloveLamp]

 

[supersonic001]

Teksun Inc on LinkedIn: #ai #technology #team #help #art #machinelearning #algorithms

Teksun Inc is proud to offer our state-of-the-art Drowsiness Detection AI Model, designed to help prevent accidents caused by driver fatigue. Our advanced…

www.linkedin.com

 

[TECH]

Maybe someone else has noticed that yesterday our Australian Patent was published 11/5/2023
84 pages including all diagrams, nice work by our brains department.

AU2022203607B1 Event-based extraction of features in a convolutional spiking neural network

6 business days then our AGM, hope you get to say what you wish to say to the right people, please try to be respectful in doing so though.
Enjoy the fellowship afterwards.

Goodnight from the West Coast......Tech

 

[Diogenese]

Maybe someone else has noticed that yesterday our Australian Patent was published 11/5/2023
84 pages including all diagrams, nice work by our brains department.

AU2022203607B1 Event-based extraction of features in a convolutional spiking neural network

6 business days then our AGM, hope you get to say what you wish to say to the right people, please try to be respectful in doing so though.
Enjoy the fellowship afterwards.

Goodnight from the West Coast......Tech

Hi Tech,

Is the first BRN patent that does not name PvdM as an inventor?

 

[Fullmoonfever]

Hadn't seen this paper by EI previously.

Says from April 2023.

Can see where we can fit in their process and also liked the exposure to the projects and developers utilising EI.

Doc here.

HERE

EDGE IMPULSE: AN MLOPS PLATFORM FOR TINY MACHINE LEARNING

Edge Impulse is a cloud-based machine learning operations (MLOps) platform for developing embedded and edge ML (TinyML) systems that can be deployed to a wide range of hardware targets. Current TinyML workflows are plagued by fragmented software stacks and heterogeneous deployment hardware, making ML model optimizations
difficult and unportable. We present Edge Impulse, a practical MLOps platform for developing TinyML systems at scale. Edge Impulse addresses these challenges and streamlines the TinyML design cycle by supporting various software and hardware optimizations to create an extensible and portable software stack for a multitude of embedded systems.

As of Oct. 2022, Edge Impulse hosts 118,185 projects from 50,953 developers.


4.5 Compression and Optimization
Many different optimization types can be used to improve the performance of ML and DSP algorithms when deployed
to edge devices. Several types of optimization are supported by Edge Impulse, either out-of-the-box or via extensibility.

The optimization areas are model compression and optimization, code optimization, and device-specific optimization.

Model compression and optimization techniques are applied either during or after training and result in models with a reduced size or computational burden when deployed to edge devices.

Compression techniques available out-of-the-box in Edge Impulse include fully int-8, weight and activation quantization (Jacob et al., 2017) and operator fusion (goo,
2022). Quantization-aware training is supported when converting a model to Brainchip’s Neuromorphic format (bus,
2022)

 

[Fullmoonfever]

Just surfing and came across this older article from late 2022 with some snips below.

Wonder if we know anyone that is in the TinyML space and do we know anyone that develops and sells MCUs and ASICs

TinyML SaaS to become a billion-dollar market by 2030

ABI Research forecasts that TinyML SaaS revenue will exceed US$220 million in 2022 and potentially become a billion-dollar market by 2030.

cfotech.asia

The TinyML Foundation, which gathers most of the prominent vendors in this space, has substantially expanded in recent years.

Similar expansion has been in the applications of TinyML; with forest fire detection, shape detection, and seizure detection among some of the most spectacular use cases.

Moreover, given how central environmental sensors are to TinyML, the possibilities are extensive.

Nonetheless, ambient sensing and audio processing remain the most common applications in TinyML, with sound architectures holding an almost 50% market share in 2022. Most of these applications employ either a microcontroller (MCU) or an Application-Specific Integrated Circuit (ASIC).

The personal and work devices sector will be the most significant increase in the near future.

“Any sensory data from an environment can probably have an ML model applied to that data. Some of the most common applications include word spotting, object recognition, object counting, and audio or voice detection,” explains David Lobina, Artificial Intelligence and Machine Learning Research Analyst at ABI Research.....

......“The role of software is crucial, and vendors must develop software tools to automate TinyML itself. Finally, new technology will be required to bring about ever more sophisticated TinyML models. Neuromorphic computing and chips, along with the corresponding technique of Spiking Neural Networks, would bode well for the future,” adds Lobina.

 

[Fullmoonfever]

Like that EI providing the tutorials to promote and make utilisation of Akida easier for Devs.

TensorFlow with the Edge Impulse Python SDK

This tutorial shows you how to use the Edge Impulse Python SDK to profile and deploy a machine learning model trained with TensorFlow and Keras.

www.edgeimpulse.com

TensorFlow with the Edge Impulse Python SDK​

MACHINE LEARNING, TINYML, ARTIFICIAL INTELLIGENCE, EMBEDDED DEVICES
Shawn Hymel
28 April 2023

The Edge Impulse Python SDK allows you to profile and deploy your machine learning (ML) models that were developed in nearly any ML framework to a number of hardware targets. Such targets include TensorFlow Lite, TensorFlow Lite Micro (TFLM), EON Compiler, vendor-specific toolchains (e.g. TensorRT), and exotic silicon (e.g. Brainchip neuromorphic processor). This SDK wraps the model with a number of useful pre- and post-processing functions along with device specific optimizations and configurations to make deployment as simple as possible....

.....For deep learning accelerators that do not support TFLM (such as neuromorphic processors) you can also convert a TF model directly into a compatible package, for example BrainChip AKD1000, Ethos-U55 microNPU, and Syntiant NDP.....

......

 

[FrederikSchack]

Hi Frederick,

1. Akida is asynchronous, responding to input events, so I don't think it can be overclocked. However, producing it as, say, 7nm will increase speed and reduce power.

2. Because Akida does not need cooling because it is so low power, it would be an ideal buried layer in a stacked chip arrangement. One potential application would be the Sony/Prophesee image sensor which uses Sony's 3D technology. I imagine the Prophesee DVS is also low power as it only fires on events. Also some IR night-vision systems need to be cooled, so Akida would be a good fit as it would reduce cooling requirements. Akida is designed for multiple chip applications.

One of my friends whom I convinced to buy BRN was asking about Ai following the recent ChatGPT publicity, so I sent the following:

You asked whether BrainChip's Akida was involved in AI, and I subsequently realized that ChatGPT has been in the news lately, not always favourably.

It is not related to ChatGPT. It could be used on the input side, but, at this stage, not on the output side generating the responses. That is all done with software using a web browser and language interpreting software.

Akida is a spiking neural network which imitates the brain's neural/synaptic processes. It does not use the same digital mathematical methods as conventional computers which consume enormous amounts of processing power in performing object recognition. Think of objects in a field of view as having a line around the edge, ie, where the pixel illumination changes. Akida compares the edge line with stored edge lines in a library of edge lines, whereas an object recognition program in a conventional computer does a pixel-by-pixel comparison of full frame images. Akida also does it in silicon, not in software. It is loaded with its compact object library data and does a hardware comparison, not a software comparison.

Akida is used to classify input signals, voice, video, etc. It does not generate replies.

It can be used for autonomous driving sensors such as LiDaR, radar, video, and event-cameras (shutterless cameras which detect only changes in the pixel illumination) aka DVS (Dynamic Vision Systems) to recognize the surroundings. It is also be used in-cabin driver monitoring, speech recognition, and any application which involves sensor signal interpretation such as vibration sensing to anticipate the need for maintenance. I suppose it could detect when a Boeing spits a turbine blade out.

It is used in Mercedes EQXX concept car which achieved 1000 km in the in-vehicle voice recognition system where it performed 5 to 10 times more efficiently that the other systems they had tested in a system where every Watt counts.

It can be used for autonomous drone navigation and image detection.

NASA are trialing it for navigation, Mars rovers, as well as wireless communication via a company called Intellisense.

Similarly, ISL (Information Systems Limited) is using it in USAF trials.

Akida is listed by ARM as being compatible with all their processors, as well as being compatible with ARM's upstart challenger SiFive which use RISC-V architecture compared with ARM's RISC-IV.

It is also available to Intel Foundry Services (IFS) customers.

It is also part of a few US university computer courses including Carnegie Mellon.

BrainChip's main business is licensing the design of Akida, a similar business model to ARM. This does limit the customer base to those who can afford the licence fee and the cost of incorporating the design in their product and manufacturing the chips. It also introduces a large time lag for designing, making and testing the chips.

A company called Renesas has licenced the Akida IC design and will be bringing out microprocessors capable of handling 30 frames per second (fps) later this year. Akida is capable of much faster fps (> 1000) but Renesas only licenced two Akida nodes out of a possible 80.

Similarly, MegaChips is also in the process of producing chips containing the Akida design.

A second generation of Akida will also be available later this year. This will have the capability to determine the speed and direction of objects in hardware rather than relying on software to process the images identified by Akida. Software is much slower and more power hungry.

Nice

With regards to point 1, I wasen't thinking overclocking as I know it's an SNN, but thinking generally that there must be ways to run it more agressively? More input? Extreme sensor fusion? More complicated models?

Another perspective could be that it may be easier to achieve 1nm if the heat dissipation is minuscle?

 

[FrederikSchack]

Great post Dio!

It's also important to zoom out and think about who wants Akida and why.
They aren't chasing extra flops (equivalent) per watt, they are trying to reduce watts per operation.
The super computers and Nordic self-cooling warehouses can handle the big stuff, but only low power, edge NM technology will solve the use-cases you mention above.

Great post again, it's easy to lose track of all the connections and reasons why Akida excels.

Thanks. I'm aware that the Edge is where we'll probably find our bread and butter first, but there must be a potential in the server market too.

If one Akida-P can challenge some of nVidias gear in some tasks and you pack 50 Akida-P on a card and still don't need cooling, then some in the server market may become curious

 

[JDelekto]

Nice

With regards to point 1, I wasen't thinking overclocking as I know it's an SNN, but thinking generally that there must be ways to run it more agressively? More input? Extreme sensor fusion? More complicated models?

Another perspective could be that it may be easier to achieve 1nm if the heat dissipation is minuscle?

Because it is event-driven, I think one of the metrics which will probably be used to stress not just Akida but other SNNs would be the timing between the pulses of the spikes.

However, I would think that by processing events more frequently, one would also be consuming more power to do so. I don't know the minimum spike distance that Akida will process comfortably. It may already be capable of keeping up with the sensors that exist.

While spikes with a short time between each pulse might be most beneficial for detecting or inferencing things in a video stream, the actual training of the network itself may not require such rapid input. However, it may require more passes in training to get better accuracy.

Memory, parameters in the model, power consumption, and cost will be factors, but Akida will also require some different benchmarking criteria than the existing AI accelerators that crunch matrices.

 

[rgupta]

Thanks. I'm aware that the Edge is where we'll probably find our bread and butter first, but there must be a potential in the server market too.

If one Akida-P can challenge some of nVidias gear in some tasks and you pack 50 Akida-P on a card and still don't need cooling, then some in the server market may become curious

I assume company is concentrating it's efforts on the edge. Opening another front will become lot more challenging.
But a nice thought process and may be we will be there as well.
Dyoor

 

[Schnitzel lover]

Yes there was some strong buying “up” of stock yesterday and today. Great to see that momentum is shifting.

So good was the buying up today that I observed there were whole SP levels bought out within seconds at some points and I am talking 300k to 400k worth of shares bought in a big chunk so that is insto’s or shorters covering and either way it’s great as it’s a sign that they know the price is cheap and it’s not staying here for very much longer.

It’s too early to tell. It’s a good start.. Look at that price action you got end of 2021.. That shape is what you want.. Staircase action..

It had to start somewhere.. That was on Mercedes Benz hype. One could expect with some positive newsflow into H2 of 2023, the market will start pricing it in.. So here’s to crossing fingers toes and everything that the corner has been turned, and there’s a lot to be looking forward to..