BRN Discussion Ongoing

[wilzy123]

Kochie spoke like he just had an aneurysm. Could hardly string any words together and really lowered the already low integrity of this podcast. It seems they let anyone/nobodies commentate on that podcast.... really inconsistent quality (i.e. ranging from incomprehensible to shit).

This is the WANCA's on the call

 

[Dozzaman1977]

On the Bosch website under new products is this sensor with AI, the use cases sound familar. Not saying akida IP is inside but i have my fingers crossed!!!!!

 

[Townyj]

Hi IDD,

I think the imperative to use Akida is diminished by each of the following:
Available mains power;
Non-time-critical;
Low data volumes and data rates.

The corollary of course being:
Battery powered;
Time critical;
Large volumes of data/high data rates;

where Akida is the best choice.

So, for things like EVs, mobile phones, laptops, wireless monitors, Akida is the most practical answer.

On the other hand, supermarket checkout stations, white goods, maybe even vibration monitors, and the like can get by without Akida. That's not to say that Akida is not the best solution for those applications, just that an inferior solution could provide a workable solution.

These things can run as software on on CPUs, GPUs, MCUs, or even using analog SNNs where accuracy is not at a premium.

That said, I hope that Akida will be so ubiquitous that it will be so cost effective as, to coin an oxymoron, to make Akida a no-brainer.

So, just at the moment, while there may be obvious use cases for Akida, until we have confirmation from BrainChip, I try to suppress my enthusiasm.

The cost of Akida would be a worth while solution over others though for supermarkets.

Totally agree with your enthusiasm, once we gain more traction who knows what we could end up in.

World Domination!!

.

 

[Bravo]

On the Bosch website under new products is this sensor with AI, the use cases sound familar. Not saying akida IP is inside but i have my fingers crossed!!!!!
View attachment 16310 View attachment 16312

What happens if the sensor detects a bad odour? Does it set off an alarm?

 

[Learning]

On the subject of LG

Is there anything here Dio?

EP3666954A1 - Washing machine - Google Patents

Disclosed herein is a washing machine (300) including a first data acquirer (310) configured to collect data related to a laundry pattern of a user, a second data acquirer (320) configured to collect data related to context information, and a processor (360) configured to provide the laundry...

patents.google.com

+
Learning.

 

[Diogenese]

On the subject of LG

Is there anything here Dio?

EP3666954A1 - Washing machine - Google Patents

Disclosed herein is a washing machine (300) including a first data acquirer (310) configured to collect data related to a laundry pattern of a user, a second data acquirer (320) configured to collect data related to context information, and a processor (360) configured to provide the laundry...

patents.google.com

View attachment 16321 +
Learning.

Doesn't seem like they had heard of SNNs back in December 2018:

• [0136] In machine learning, a number of machine learning algorithms for classifying data have been developed. Decision tree, Bayesian network, support vector machine (SVM), and artificial neural network (ANN) are representative examples of the machine learning algorithms.

[0141] The ANN is a model used for machine learning and is a statistical learning algorithm inspired from a neural network (for example, brains in a central nervous system of animals) of biology in machine learning and cognitive science.


[0145] The ANN may include network models such as a deep neural network (DNN), a recurrent neural network (RNN), a bidirectional recurrent deep neural network (BRDNN), a multilayer perceptron (MLP), and a convolutional neural network (CNN), but is not limited thereto.

 

[Learning]

Doesn't seem like they had heard of SNNs back in December 2018:

• [0136] In machine learning, a number of machine learning algorithms for classifying data have been developed. Decision tree, Bayesian network, support vector machine (SVM), and artificial neural network (ANN) are representative examples of the machine learning algorithms.

[0141] The ANN is a model used for machine learning and is a statistical learning algorithm inspired from a neural network (for example, brains in a central nervous system of animals) of biology in machine learning and cognitive science.


[0145] The ANN may include network models such as a deep neural network (DNN), a recurrent neural network (RNN), a bidirectional recurrent deep neural network (BRDNN), a multilayer perceptron (MLP), and a convolutional neural network (CNN), but is not limited thereto.

Thanks Dio,

Algorithms.

Learning

 

[Diogenese]

Thanks Dio,

Algorithms.

Learning

That's right.

"all representative of machine learning algorithms";

"a model for machine learning";

"network models such as a deep neural network (DNN), a recurrent neural network (RNN), a bidirectional recurrent deep neural network (BRDNN), a multilayer perceptron (MLP), and a convolutional neural network (CNN)"
in 2018 would all have been software solutions.

They are talking about software (CPU programs).

 

[equanimous]

That's right.

"all representative of machine learning algorithms";

"a model for machine learning";

"network models such as a deep neural network (DNN), a recurrent neural network (RNN), a bidirectional recurrent deep neural network (BRDNN), a multilayer perceptron (MLP), and a convolutional neural network (CNN)"
in 2018 would all have been software solutions.

They are talking about software (CPU programs).

Hi Diogenese,

Im curious from your point of view, what are you most excited about with Akida and in what applications?

Is there a use case that hasnt been seen to date which you think could really benefit from Akida?

Do you have any thoughts on future editions of Akida and its potential?

Thanks in advance

 

[Steve10]

FYI - MegaChips invests in Australian semiconductor company.

Sydney-based semiconductor company Morse Micro raises $140 million to progress Wi-Fi technology​

MORGANNE KOPITTKE
SEPTEMBER 9, 2022

Sydney-based fabless semiconductor company Morse Micro is ushering in a new era for its Wi-Fi HaLow technology and plans to use the $140 million in Series B funding to revolutionise Australia’s digital future.

Leading Japan-based ASIC and SoC services company, MegaChips Corporation, led the funding round, with participation from existing investors including Blackbird Ventures, Main Sequence Ventures, Clean Energy Finance Corporation, Skip Capital, Uniseed, SpringCapital, and Malcolm and Lucy Turnbull among others.

The capital raised will be used to achieve scale and demand for Morse Micro’s Wi-Fi HaLow technology, focusing on the design of new solutions, while accelerating the go-to-market strategy for its existing Wi-Fi HaLow chips and modules.

Morse Micro claims its Wi-Fi HaLow “is the first Wi-Fi standard tailored to meet the needs of the IoT, making it a superior alternative to traditional Wi-Fi alternatives and many other wireless technologies”.

Launched in 2016, Morse Micro was founded by Wi-Fi pioneers and innovators, Michael De Nil and Andrew Terry, joined by the original Wi-Fi inventor Professor Neil Weste and wireless industry veterans, whose teams designed Wi-Fi chips for billions of smartphones.

Morse Micro co-founder and CEO Michael De Nil says the company is now at “an exciting inflection point”.

“With MegaChips’ financial backing and robust manufacturing and sales support, Morse Micro will be poised to achieve our goal of revolutionising IoT connectivity with our growing portfolio of Wi-Fi HaLow SoCs, modules, software and development tools,” he said.

“MegaChips shares our vision to revolutionise connectivity and build enduring Wi-Fi HaLow solutions for the future. We are delighted to partner with them, and for the support we’ve had from all of our investors, as we take the next step forward in our company’s journey toward market scale and leadership.”

MegaChips group president and CEO Tetsuo Hikawa says Morse Micro’s leadership team has already made a “tremendous impact” on the Wi-Fi industry.

“MegaChips pursues strategic partnerships and investments with promising startups like Morse Micro that have innovative ideas, cutting-edge applications and are expanding their businesses in areas such as industrial IoT, wireless communications and energy control,” he said.

The multimillion-dollar investment also brings with it a strategic business partnership between Morse Micro and MegaChips, which will see the two companies engage in joint sales and promotion activities to grow the market for Wi-Fi HaLow solutions.

In addition to manufacturing Morse Micro’s IEEE 802.11ah compliant semiconductors and modules, MegaChips will provide sales support and new distribution channels that will deliver scale across the region.

 

[Diogenese]

Hi Diogenese,

Im curious from your point of view, what are you most excited about with Akida and in what applications?

Is there a use case that hasnt been seen to date which you think could really benefit from Akida?

Do you have any thoughts on future editions of Akida and its potential?

Thanks in advance

Hi equanimous,
There's quite a lot to be excited about:

ARM, SiFive, Valeo, nViso, Prophesee, Edge Impulse, MegaChips, Mercedes ...

Valeo LiDaR is one of the most exciting in that, combined with the EV surge in the real world and EU regulations for green tech and driver monitoring, and with AI being flavour of the month, it feels like we are surfing two waves at the same time. Valeo are the leading (sole) supplier of LiDaR ... and LiDaR is a sweet spot for Akida.

We signed a JD agreement with Valeo 2 years ago, so, with any luck, there'll be some road trials early next year if not some time this year with Scala 3.

nViso gives us another iron in the EV fire.

Edge Impulse can facilitate the matching of any task to Akida.

MegaChips gives us access to everything else.

Prophesee is another product (DVS) which is uniquely adapted for Akida because it generates spikes straight from the oven.

SiFive is ambitious to take a large chunk of ARM's business, and we have 5 bob each way.

I've left Renesas out as they are only licensed for 2 nodes, apparently preferring their own DRP for the heavy lifting.

Leaving aside mobile phones, laptops, cloud servers, I'm sure there will be many applications which have not yet surfaced but my crystal ball does not have the capacity to list them all. There are all sorts of possibilities - large computer makers for one. I hope to see the day when all processors (CPU, GPU, MCU) will incorporate Akida SNN IP.

Future versions of Akida will have LSTM (useful for predicting the motion of objects), and PvdM is working on cortical neuron bundles, but that's above my pay grade.

The nervous system is full duplex - there is an inward path carrying information from our senses to the brain, and there is a separate path carrying instructions from our brain to our muscles. At present, Akida only deals with the inbound signals. I guess the outward bound signals are also spikes, but at present, we rely on a CPU to send instructions. However, the instructions are probably small beer compared with the sensor data.

 

[Diogenese]

Hi equanimous,
There's quite a lot to be excited about:

ARM, SiFive, Valeo, nViso, Prophesee, Edge Impulse, MegaChips, Mercedes ...

Valeo LiDaR is one of the most exciting in that, combined with the EV surge in the real world and EU regulations for green tech and driver monitoring, it feels like we are surfing two waves at the same time. Valeo are the leading (sole) supplier of LiDaR ... and LiDaR is a sweet spot for Akida.

We signed a JD agreement with Valeo 2 years ago, so, with any luck, there'll be some road trials early next year if not some time this year with Scala 3.

nViso gives us another iron in the EV fire.

Edge Impulse can facilitate the matching of any task to Akida.

MegaChips gives us access to everything else.

Prophesee is another product (DVS) which is uniquely adapted for Akida because it generates spikes straight from the oven.

SiFive is ambitious to take a large chunk of ARM's business, and we have 5 bob each way.

I've left Renesas out as they are only licensed for 2 nodes, apparently preferring their own DRP for the heavy lifting.

Leaving aside mobile phones, laptops, cloud servers, I'm sure there will be many applications which have not yet surfaced but my crystal ball does not have the capacity to list them all. There are all sorts of possibilities - large computer makers for one. I hope to see the day when all processors (CPU, GPU, MCU) will incorporate Akida SNN IP.

Future versions of Akida will have LSTM (useful for predicting the motion of objects), and PvdM is working on cortical neuron bundles, but that's above my pay grade.

The nervous system is full duplex - there is an inward path carrying information from our senses to the brain, and there is a separate path carrying instructions from our brain to our muscles. At present, Akida only deals with the inbound signals. I guess the outward bound signals are also spikes, but at present, we rely on a CPU to send instructions. However, the instructions are probably small beer compared with the sensor data.

I meant to add that Scala 3 may need the LSTM version of Akida. I'm pretty sure we would have the software simulation of this, but I don't know if we've made a FPGA, or if Valeo have included the IP in a SoC.

 

[Learning]

Thank you equanimous, for asking the right questions .

Thank you Dio, for your answers.

Live and learn.
Always Learning.

 

[Stable Genius]

DeepPicarMicro Crams NVIDIA's PilotNet Autonomous Vehicle Neural Network Into a Raspberry Pi Pico​

DeepPicarMicro Crams NVIDIA's PilotNet Autonomous Vehicle Neural Network Into a Raspberry Pi Pico - Hackster.io

www-hackster-io.cdn.ampproject.org

 

[equanimous]

Thank you equanimous, for asking the right questions .

Thank you Dio, for your answers.

Live and learn.
Always Learning.

Ill pose the questions for everyone here now as It would be interesting to hear everyones response.

Im curious from your point of view, what are you most excited about with Akida and in what applications?

Is there a use case that hasnt been seen to date which you think could really benefit from Akida?

Do you have any thoughts on future editions of Akida and its potential?

 

[stuart888]

Hi equanimous,
There's quite a lot to be excited about:

ARM, SiFive, Valeo, nViso, Prophesee, Edge Impulse, MegaChips, Mercedes ...

Valeo LiDaR is one of the most exciting in that, combined with the EV surge in the real world and EU regulations for green tech and driver monitoring, and with AI being flavour of the month, it feels like we are surfing two waves at the same time. Valeo are the leading (sole) supplier of LiDaR ... and LiDaR is a sweet spot for Akida.

We signed a JD agreement with Valeo 2 years ago, so, with any luck, there'll be some road trials early next year if not some time this year with Scala 3.

nViso gives us another iron in the EV fire.

Edge Impulse can facilitate the matching of any task to Akida.

MegaChips gives us access to everything else.

Prophesee is another product (DVS) which is uniquely adapted for Akida because it generates spikes straight from the oven.

SiFive is ambitious to take a large chunk of ARM's business, and we have 5 bob each way.

I've left Renesas out as they are only licensed for 2 nodes, apparently preferring their own DRP for the heavy lifting.

Leaving aside mobile phones, laptops, cloud servers, I'm sure there will be many applications which have not yet surfaced but my crystal ball does not have the capacity to list them all. There are all sorts of possibilities - large computer makers for one. I hope to see the day when all processors (CPU, GPU, MCU) will incorporate Akida SNN IP.

Future versions of Akida will have LSTM (useful for predicting the motion of objects), and PvdM is working on cortical neuron bundles, but that's above my pay grade.

The nervous system is full duplex - there is an inward path carrying information from our senses to the brain, and there is a separate path carrying instructions from our brain to our muscles. At present, Akida only deals with the inbound signals. I guess the outward bound signals are also spikes, but at present, we rely on a CPU to send instructions. However, the instructions are probably small beer compared with the sensor data.

Dio takes the role of the Activation Function, perfectly summarizing the tech situation brilliantly. Must read by all.
Fantastic post @Diogenese!

 

[stuart888]

Love all the Enabling Tech that is part of the Akida workflow, including the sensors!

The Best Sensors 2022 Awards was loaded with lots of interesting nominees and winners.

All the enabling tech is blasting off too, really good for sensor agnostic Brainchip.

 

[Learning]

Ill pose the questions for everyone here now as It would be interesting to hear everyones response.

Im curious from your point of view, what are you most excited about with Akida and in what applications?

Is there a use case that hasnt been seen to date which you think could really benefit from Akida?

Do you have any thoughts on future editions of Akida and its potential?

Hi equanimous,

As I have said in the past, I don't have an It, engineering or science back ground to answer the technical aspects of Akida.

However, as a shareholder, I have tried to learn what Brainchip and Akida is about and why I should invest in it.

I can only go off from the the information Brainchip had released, and learnt from what other info shareholders had generously shared.

I am excited because of Akida ability to 'one shot learning', as learning would be a cognitive ability where we human improve with practice overtime. Now Akida can mimics that ability, then the use case would be diverse.

I take an example of Mercedes "a car that thinks like you" I would like to think that when Akida is operational in the Mercedes, When the owner first time driving it, it will learn of the face, height to adjust the seat, music & volume preference, all the comforts setting thats we has to adjust if the car is co-share. Although they are simply things, but from a cognitive aspect it has to be learnt over time.

With the idea of inference, some technical minded would think differently. However, my take on it is. With Akida ability to learn overtime, eventually it's has learnt enough data to make small judgement( drawing inference) to conclude or make decisions within the chip/or operation.

Learning.
It's great to be a shareholder.

 

[Dhm]

FYI - MegaChips invests in Australian semiconductor company.

Sydney-based semiconductor company Morse Micro raises $140 million to progress Wi-Fi technology​

MORGANNE KOPITTKE
SEPTEMBER 9, 2022

Sydney-based fabless semiconductor company Morse Micro is ushering in a new era for its Wi-Fi HaLow technology and plans to use the $140 million in Series B funding to revolutionise Australia’s digital future.

Leading Japan-based ASIC and SoC services company, MegaChips Corporation, led the funding round, with participation from existing investors including Blackbird Ventures, Main Sequence Ventures, Clean Energy Finance Corporation, Skip Capital, Uniseed, SpringCapital, and Malcolm and Lucy Turnbull among others.

The capital raised will be used to achieve scale and demand for Morse Micro’s Wi-Fi HaLow technology, focusing on the design of new solutions, while accelerating the go-to-market strategy for its existing Wi-Fi HaLow chips and modules.

Morse Micro claims its Wi-Fi HaLow “is the first Wi-Fi standard tailored to meet the needs of the IoT, making it a superior alternative to traditional Wi-Fi alternatives and many other wireless technologies”.

Launched in 2016, Morse Micro was founded by Wi-Fi pioneers and innovators, Michael De Nil and Andrew Terry, joined by the original Wi-Fi inventor Professor Neil Weste and wireless industry veterans, whose teams designed Wi-Fi chips for billions of smartphones.

Morse Micro co-founder and CEO Michael De Nil says the company is now at “an exciting inflection point”.

“With MegaChips’ financial backing and robust manufacturing and sales support, Morse Micro will be poised to achieve our goal of revolutionising IoT connectivity with our growing portfolio of Wi-Fi HaLow SoCs, modules, software and development tools,” he said.

“MegaChips shares our vision to revolutionise connectivity and build enduring Wi-Fi HaLow solutions for the future. We are delighted to partner with them, and for the support we’ve had from all of our investors, as we take the next step forward in our company’s journey toward market scale and leadership.”

MegaChips group president and CEO Tetsuo Hikawa says Morse Micro’s leadership team has already made a “tremendous impact” on the Wi-Fi industry.

“MegaChips pursues strategic partnerships and investments with promising startups like Morse Micro that have innovative ideas, cutting-edge applications and are expanding their businesses in areas such as industrial IoT, wireless communications and energy control,” he said.

The multimillion-dollar investment also brings with it a strategic business partnership between Morse Micro and MegaChips, which will see the two companies engage in joint sales and promotion activities to grow the market for Wi-Fi HaLow solutions.

In addition to manufacturing Morse Micro’s IEEE 802.11ah compliant semiconductors and modules, MegaChips will provide sales support and new distribution channels that will deliver scale across the region.

I am trying to decide if Morse Micro is a friend or a foe. It's big sell - if I understand it - is it creates an expanded Wi-Fi network for applications that otherwise couldn't link up with each other, like a perimeter security system. They are using similar phrases and buzzwords to us like ultra low power and security.

Comments and opinions welcome.

https://www.morsemicro.com

 

[Quatrojos]

Ill pose the questions for everyone here now as It would be interesting to hear everyones response.

Im curious from your point of view, what are you most excited about with Akida and in what applications?

Is there a use case that hasnt been seen to date which you think could really benefit from Akida?

Do you have any thoughts on future editions of Akida and its potential?

I would’ve loved a baby monitor that only woke me when my babies were distressed…