BRN Discussion Ongoing

[Fullmoonfever]

Short Yole write up from Apr which reads quite positively.

MegaChips focuses on edge AI with custom ASIC solutions

Japan’s largest ASIC company expands to U.S. market. MegaChips, the leading custom ASIC company in Japan, today announced the launch of its AI Partner Program, which allows companies to integrate powerful AI capabilities without requiring in-house AI experts, allowing vendors to focus on their...

www.yolegroup.com

MegaChips focuses on edge AI with custom ASIC solutions​

APRIL 22, 2022

Japan’s largest ASIC company expands to U.S. market. MegaChips, the leading custom ASIC company in Japan, today announced the launch of its AI Partner Program, which allows companies to integrate powerful AI capabilities without requiring in-house AI experts, allowing vendors to focus on their key strengths and ensuring top quality for the final product.
The AI Partner Program marks the entry of MegaChips into the global Edge AI chips market, which was valued at $9 billion in 2020, and is projected to reach $59.6 billion by 2030 – an average growth rate of 21.2%.
“The AI chip industry is going through many changes, including a pivot from a saturated data center market to emerging use cases for integrated processors, such as the ones you’d find in smart devices”, asserts Adrien Sanchez, Technology & Market Analyst, Computing at Yole Développement (Yole). “Edge AI chips benefit companies by allowing them to analyze data from connected devices without sending massive amounts of data into the cloud, which often results in massive costs and potential security risks.” (1)
For systems companies, some benefits of the MegaChips AI Partner Program include a dedicated team of engineers that work collaboratively with customers to identify the best ways to implement desired AI functionalities, custom “proof of concept” demonstrations, and optimization strategy in context of a complete system. For IP and ASSP vendors, MegaChips eliminates the need for hiring in-house back-end chip implementation teams.
MegaChips is also announcing its expansion into the U.S. market after extensive success in Japan. MegaChips is now delivering its full-service ASIC solution in the U.S. and offering off-the-shelf access to industry-standard IP components and secure, inhouse design services along with full manufacturing support.
“MegaChips is thrilled to offer the most turn-key solution for enterprise companies looking to implement AI technology,” said Douglas Fairbairn, Director of Business Development. “The expansion to the United States is an excellent opportunity for us to bring our edge AI expertise to some of the most innovative technology companies. Be it sensing, voice and image recognition, or other applications, MegaChips is the first and best choice for implementing Edge AI from ideation to silicon

 

[Fact Finder]

I think you need to recheck your figures and post back more accurate figures based on the last half yearly statement

I could not stand the suspense so I took off my shoes and socks and guessed a bit but 4. a,b & c add up to approximately $10 million and I multiplied that by two in my head so cannot show the working but it came to $20 million. To check my workings I went to the last Annual Report produced below showing in bold the total then was $21,234,175 which included all those extra expenses with Socionext and TSMC to make that little plastic & metal thing that everyone is so obsessed with here. Of course Brainchip has cut out this expense going forward as they are now IP only.

From the Annual Report:
REVIEW OF OPERATIONS
The financial results of the Group are presented in US dollars, unless otherwise referenced. Although the pandemic remains an ongoing global challenge, there has been minimal impact on our business other than certain restrictions when meeting and engaging with our customers face-to-face, and the supply of certain core components used in the construction of development kits.

Overview
The Group made a net loss after income tax for the year ended 31 December 2021 of $20,981,309 (2020: $26,822,049). The current year loss included non-cash losses of $1,375,950 resulting from the fair valuation of the LDA financial liabilities (derivatives) recognised due to the agreed pricing mechanism and the put option premium. Revenues for the year ended 31 December 2021 of $1,588,483 increased 1,215% from $120,829 in 2020. The Company executed a license contract in December 2021 as well as multiple Early Access Programme (“EAP”) product sales and engineering support services agreements during the year. Directors’ Report BrainChip Holdings Ltd 2021 Annual Report

5 REVIEW OF OPERATIONS (Continued)
Total operating expenses for the year ended 31 December 2021 of $21,234,175 increased 89% from $11,242,032 incurred in the year ended 31 December 2020.
This increase was attributable to:
1) Research & development (R&D) expenses of $9,133,213 for the current period increased 77%, or $3,980,974 from a year ago. R&D costs comprise employee expense, contractor and other research and development costs, and amortisation of capitalised R&D intangible assets. Movements in R&D costs are summarised as follows: a) 41% increase in employee expenses reflecting the expansion of headcount into both Australia (2 additional) and India (5 additional) plus merit increases to staff who supported the achievements of the Company throughout 2020 and 2021, offset by increased credits received or receivable from government authorities; and b) recognition of $4,067,000 of third-party pre-development services being the final progressive payments to Socionext related to the fabrication of the AkidaTM device (2020: $1,050,000);
2) Selling & marketing (S&M) expenses of $2,486,927 for the current period increased 74%, or $1,060,426 from a year ago. The increase reflects management’s focus on the promotion and marketing of product, and the targeting of potential customers through the engagement of external marketing consultants and expansion of the S&M headcount;
3) General & administrative (G&A) expenses of $5,251,834 for the current period increased 63% overall, or $2,024,187 from the same period a year ago as a result of: a) the appointment of a new CEO in November 2021 as well as the termination expense of the previous CEO effective from March 2021, plus additional headcount in the administrative team; b) an increase in directors fee expense as a result of a full year of fees in 2021 with an additional board member appointed; c) increased legal and professional costs supporting employment and selling functions, and d) increased listing related expenses as a result of increased shareholder activities during the current year; and
4) Share-based payment expense of $4,362,201 for the current period increased 204%, or $2,926,556 from the same period a year ago. Share-based payments expense is non-cash in nature, and represents the current period vesting expense for options, restricted stock units and performance rights issued to directors, employees and consultants, offset by the value of options that have been forfeited during the year.

From the recent Half Yearly Report

7. LOSS PER SHARE 30 June 2022 30 June 2021 US$ US$ Net loss attributable to ordinary shareholders for basic and diluted earnings per share (8,255,802) (9,298,244) US cents per share US cents per share Basic and diluted loss per share (0.46) [2022 ~v~ 2021] (0.55)

It is also interesting that the attributed net loss per share fell to 46 cents over the last half yearly report when it was 55 cents per share.

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Deena]

I can't wait until they have fully automated buses. And I've got to tell you it doesn't always pay to do good deeds.

I gave up my seat on the bus the other day for a blind man (blind Freddy?). Anyway that's how I lost my job ... as a bus driver.

 

[Dozzaman1977]

I could not stand the suspense so I took off my shoes and socks and guessed a bit but 4. a,b & c add up to approximately $10 million and I multiplied that by two in my head so cannot show the working but it came to $20 million. To check my workings I went to the last Annual Report produced below showing in bold the total then was $21,234,175 which included all those extra expenses with Socionext and TSMC to make that little plastic & metal thing that everyone is so obsessed with here. Of course Brainchip has cut out this expense going forward as they are now IP only.

From the Annual Report:
REVIEW OF OPERATIONS
The financial results of the Group are presented in US dollars, unless otherwise referenced. Although the pandemic remains an ongoing global challenge, there has been minimal impact on our business other than certain restrictions when meeting and engaging with our customers face-to-face, and the supply of certain core components used in the construction of development kits.

Overview
The Group made a net loss after income tax for the year ended 31 December 2021 of $20,981,309 (2020: $26,822,049). The current year loss included non-cash losses of $1,375,950 resulting from the fair valuation of the LDA financial liabilities (derivatives) recognised due to the agreed pricing mechanism and the put option premium. Revenues for the year ended 31 December 2021 of $1,588,483 increased 1,215% from $120,829 in 2020. The Company executed a license contract in December 2021 as well as multiple Early Access Programme (“EAP”) product sales and engineering support services agreements during the year. Directors’ Report BrainChip Holdings Ltd 2021 Annual Report

5 REVIEW OF OPERATIONS (Continued)
Total operating expenses for the year ended 31 December 2021 of $21,234,175 increased 89% from $11,242,032 incurred in the year ended 31 December 2020.
This increase was attributable to:
1) Research & development (R&D) expenses of $9,133,213 for the current period increased 77%, or $3,980,974 from a year ago. R&D costs comprise employee expense, contractor and other research and development costs, and amortisation of capitalised R&D intangible assets. Movements in R&D costs are summarised as follows: a) 41% increase in employee expenses reflecting the expansion of headcount into both Australia (2 additional) and India (5 additional) plus merit increases to staff who supported the achievements of the Company throughout 2020 and 2021, offset by increased credits received or receivable from government authorities; and b) recognition of $4,067,000 of third-party pre-development services being the final progressive payments to Socionext related to the fabrication of the AkidaTM device (2020: $1,050,000);
2) Selling & marketing (S&M) expenses of $2,486,927 for the current period increased 74%, or $1,060,426 from a year ago. The increase reflects management’s focus on the promotion and marketing of product, and the targeting of potential customers through the engagement of external marketing consultants and expansion of the S&M headcount;
3) General & administrative (G&A) expenses of $5,251,834 for the current period increased 63% overall, or $2,024,187 from the same period a year ago as a result of: a) the appointment of a new CEO in November 2021 as well as the termination expense of the previous CEO effective from March 2021, plus additional headcount in the administrative team; b) an increase in directors fee expense as a result of a full year of fees in 2021 with an additional board member appointed; c) increased legal and professional costs supporting employment and selling functions, and d) increased listing related expenses as a result of increased shareholder activities during the current year; and
4) Share-based payment expense of $4,362,201 for the current period increased 204%, or $2,926,556 from the same period a year ago. Share-based payments expense is non-cash in nature, and represents the current period vesting expense for options, restricted stock units and performance rights issued to directors, employees and consultants, offset by the value of options that have been forfeited during the year.

From the recent Half Yearly Report

7. LOSS PER SHARE 30 June 2022 30 June 2021 US$ US$ Net loss attributable to ordinary shareholders for basic and diluted earnings per share (8,255,802) (9,298,244) US cents per share US cents per share Basic and diluted loss per share (0.46) [2022 ~v~ 2021] (0.55)

It is also interesting that the attributed net loss per share fell to 46 cents over the last half yearly report when it was 55 cents per share.

My opinion only DYOR
FF

AKIDA BALLISTA

Thanks. I would of just let him try and work it out himself.
It's good to have you back posting FF.

 

[Violin1]

Kock Knock

Who's there?

A little old lady

A little old lady who?

I had no idea you could yodel


View attachment 15074

Bravo - I shake my head, really I do. Must run off and tell my grandkids.....lol

 

[chapman89]

Tesla really needs akida, i believe ARM, Valeo Mercedes and a few others are truly planning a revolution with akida and it’s capabilities.

“Tesla’s major deployment of so-called Full Self-Driving (FSD) technology is one of the most dangerous and irresponsible actions by a car company in decades,” Nader said in a statement about the autonomous system on his website.
“Together we need to send an urgent message to the casualty-minded regulators that Americans must not be test dummies for a powerful, high-profile corporation and its celebrity CEO. No-one is above the laws of manslaughter.”

Legendary US road safety campaigner Ralph Nader slams Tesla's autonomous tech

US safety campaigner Ralph Nader describes Tesla self-driving technology as "One of the most dangerous and irresponsible actions by a car company in decades"

www.drive.com.au

 

[Fact Finder]

This could prove an interesting side research area for the 1,000 Eyes:

Learner Community

nasacentral.force.com

My opinion only DYOR
FF

AKIDA BALLISTA

 

[Sirod69]

Arm is about to make its debut at CFMS2022 to discuss the cornerstones of future cloud edge digital storage and computing innovations!

On September 20, 2022, China Flash Memory Market Summit (CFMS2022) will be held at InterContinental Shenzhen OCT Samsung, Kioxia, Micron, YMTC, Solidigm, Arm, Intel, Huirong Technology, Phison, Lianyun Technology, Jiangbo Dragon Electronics, Innodisk International, Memblaze and many other large companies have communicated extensively with you on the ground and directly attacked storage technology innovation and application development.

In addition to bringing together the upstream and downstream core companies in the storage industry chain, another highlight of this summit is the architecture technology that is defining the future of computing - Arm will also be attending. Arm's high-performance, low-power processor designs and software platforms are used in advanced computing on more than 230 billion chips, and its technology safely powers electronic devices from sensors to smartphones to supercomputing, laying the foundation for trust in the digital world of chips to the cloud in the field of cybersecurity, including the provision of technology for storing and processing the huge amounts of data generated for this purpose, stimulates innovation.

In terms of digital storage, Arm has always been committed to promoting the cutting-edge development of the storage industry, including improving performance, responsiveness, energy efficiency and security.

At CFMS2022, Arm will deliver a keynote titled "Arm: The Cornerstone of Future Cloud-Edge Digital Storage and Computing Innovation" where he will share his views and insights on industry trends, including how new technologies will transform the existing layout , and the capabilities of

Arm will partner to provide the best storage and storage solutions for computing.​

Registration for China Flash Market Summit CFMS2022 has been opened, and you can register for the conference through the summit official website (https://cfms.chinaflashmarket.com/) or "Flash Market" APP.

 

[Esq.111]

Evening Chippers,

Had a quick conversation with Boardroom Pty Ltd today ,

Total Shareholders on 23/6/2022,
47,145 holders.

Total Shareholders today 26/8/2022,
46,887 holders.

= 258 less holders on the registry.

Purely contacted Boardroom to locate my position in the jungle, so before anyone asks , these two points of interest were all I received.

Needless to say eight new people / entitys now have a larger holding than myself which is great news .
Existing / new holders have been accumulating.

Great research all, always top shelf.

Regards,
Esq.

 

[MDhere]

Hi All,

The following is my personal opinion only.

Regarding the aggregate short of close to 90 Million shares of BRN.

As being mention, that would require around 7days of buying to cover. Imagine what price they will have to pay.

So,I believe the plans for the short attack is to drive the SP down and reducing the market capital of Brainchip, to force Brianchip out of the ASX 200. What happen when Brainchip is force out of the ASX 200 = Index funds will be selling BRN at the rebalance = Short cover without driving the SP up.

As of today Brainchip is Rank

229

BRN

Brainchip Holdings Ltd

$0.945

Information Technology

$1.62 B

From what we know, Brainchip is in a strong financial position and growing. From yesterday Half Year Financial Report shown $ 4.8 Million in Revenue. The short can just lose hope for Brainchip do any capital rise.

All that could be changes and trap the shorts. If general market environment are better, Brainchip have a substantial material announcement or one of our EAP "do a Merc", how fun would that be.

Anyway that's just me thinking out loud.

Learning.

I like that phrase doin a merc. very catchy! So it can "do a Merc" all the way to $5 on the next one

 

[Makeme 2020]

Evening Chippers,

Had a quick conversation with Boardroom Pty Ltd today ,

Total Shareholders on 23/6/2022,
47,145 holders.

Total Shareholders today 26/8/2022,
46,887 holders.

= 258 less holders on the registry.

Purely contacted Boardroom to locate my position in the jungle, so before anyone asks , these two points of interest were all I received.

Needless to say eight new people / entitys now have a larger holding than myself which is great news .
Existing / new holders have been accumulating.

Great research all, always top shelf.

Regards,
Esq.

Nice to see you back Buddy..........

 

[TECH]

This saying/statement is a real reflection of Brainchip's journey to date, though you could say the same about Apple, Microsoft,
Smartphones...etc...etc.

Simply put, Amara’s Law states, “Most people overestimate what they can achieve in a year and underestimate what they can achieve in ten years.”

We have seen this pattern with multiple emerging technologies. As new technology becomes a reality, it goes through an early adoption phase to find its ideal placement in our world, then grows exponentially.

Seven years in and we are on course to repeat history, what was the share price predictions for Christmas 2022 ? $2.75 plus ?

One thing I do know, while the current share price is a long way from 3.1c I personally thought that we'd be north or well north of a dollar
at this point, but it's nice to know that only 50% capital gain needs to be declared currently, which indicates that I haven't traded or offloaded around 99% of my Brainchip shares over the last 14 months or so.

I wonder why ??

Thanks to all the great posters on this site, I sense a little upswing in the mood currently...good to see Tech x

 

[Perhaps]

Have a look at the new Fujifilm flagship and its new subject recognition and tracking AF deep learning abilities. Of course in offline mode. Sounds like one shot learning, could this be the missing link to MegaChips? As it is known that MegaChips does product development for one of the big Japanese cameramakers, it's not too far away.

Fujifilm X-H2S review

The Fujifilm X-H2S is the company's most ambitious APS-C camera, using a 26MP Stacked CMOS sensor to deliver the fastest shooting, best autofocus and most extensive video specs of any X-series camera yet. We tell you what you need to know in our full review.

www.dpreview.com

This review is in German only, sorry. Here's the part in English which takes attention:

But what's really new about autofocus is precise object-oriented scene recognition. The algorithm was developed using deep learning technology and recognizes and tracks faces and eyes, but also animals, birds, cars, motorcycles and other subjects such as ships. We tested this in portrait shooting and with city pigeons - both worked perfectly. The camera recognized and tracked the subjects unerringly and without difficulty.

https://www.chip.de/test/Fujifilm-X-H2S-im-Test_184401275.html

Just to complete it, the manual of the camera:

Specifications

A website for FUJIFILM digital camera manuals.

fujifilm-dsc.com

 

[clip]

Have a look at the new Fujifilm flagship and its new subject recognition and following AF deep learning abilities. Of course in offline mode. Sounds like one shot learning, could this be the missing link to MegaChips? As it is known that MegaChips does product development for one of the big Japanese cameramakers, I'm waiting for something like this for a while.

Fujifilm X-H2S review

The Fujifilm X-H2S is the company's most ambitious APS-C camera, using a 26MP Stacked CMOS sensor to deliver the fastest shooting, best autofocus and most extensive video specs of any X-series camera yet. We tell you what you need to know in our full review.

www.dpreview.com

This review is in German only, sorry. Here's the part in English which takes attention:

But what's really new about autofocus is precise object-oriented scene recognition. The algorithm was developed using deep learning technology and recognizes and tracks faces and eyes, but also animals, birds, cars, motorcycles and other subjects such as ships. We tested this in portrait shooting and with city pigeons - both worked perfectly. The camera recognized and tracked the subjects unerringly and without difficulty.

https://www.chip.de/test/Fujifilm-X-H2S-im-Test_184401275.html

Hi Perhaps,

would be really nice. However two things keep me skeptical.
For Example the menu, where you have to switch through the different recognition modes. I would guess, with a camera with akida inside, you wouldn't need that, because it will do that automatically.

And this part in the dpreview:

We couldn't fully test them all but while animal and bird recognition modes could work very well, they seemed to need the subject to be at a specific angle to the camera, eg: perfectly side-on or facing the camera. If the subject was looking or moving away from you, the camera would often lose it.

Edit:
From Fujifilm youtube channel:

 

[Perhaps]

Digged deep into very old news related to MegaChips. This one at least gives a slight connection between MegaChips and Fujifilm. Very interesting editor notes, might give some ideas on possible MegaChips customers, no matter this is very old.

IP

Cadence helps you choose the right IP solution and capture its full value in your SoC design.

ip.cadence.com

 

[Deadpool]

Have a look at the new Fujifilm flagship and its new subject recognition and following AF deep learning abilities. Of course in offline mode. Sounds like one shot learning, could this be the missing link to MegaChips? As it is known that MegaChips does product development for one of the big Japanese cameramakers, I'm waiting for something like this for a while.

Fujifilm X-H2S review

The Fujifilm X-H2S is the company's most ambitious APS-C camera, using a 26MP Stacked CMOS sensor to deliver the fastest shooting, best autofocus and most extensive video specs of any X-series camera yet. We tell you what you need to know in our full review.

www.dpreview.com

Hey @Perhaps, very impressive piece of camera tech, Can't wait to purchase one. I hope your right, that Akida is impregnated in the architecture of this beast, but even if its not, its just fantastic how tech is evolving exponentially. Makes me really wonder what is capable in our BRN ecosystem of collaborators, and what extraordinary enterprises they are really up to?

I'm certain, (that in the not to distant future) looks so incredibly colossally bright, for not only our BRN ecosystem of shareholders, but for the whole of humanity from the proceeds of beneficial AI, that everyone will be smiling and needing to be wearing

24/7.

Its not only a tremendous feeling of wonder to be a BRN shareholder, but also an extremely humbling understanding of what is at play here.
Best regards
PS
I sincerely think that Sean is going to deliver us to the promised land

 

[Slymeat]

Arm is about to make its debut at CFMS2022 to discuss the cornerstones of future cloud edge digital storage and computing innovations!

On September 20, 2022, China Flash Memory Market Summit (CFMS2022) will be held at InterContinental Shenzhen OCT Samsung, Kioxia, Micron, YMTC, Solidigm, Arm, Intel, Huirong Technology, Phison, Lianyun Technology, Jiangbo Dragon Electronics, Innodisk International, Memblaze and many other large companies have communicated extensively with you on the ground and directly attacked storage technology innovation and application development.

In addition to bringing together the upstream and downstream core companies in the storage industry chain, another highlight of this summit is the architecture technology that is defining the future of computing - Arm will also be attending. Arm's high-performance, low-power processor designs and software platforms are used in advanced computing on more than 230 billion chips, and its technology safely powers electronic devices from sensors to smartphones to supercomputing, laying the foundation for trust in the digital world of chips to the cloud in the field of cybersecurity, including the provision of technology for storing and processing the huge amounts of data generated for this purpose, stimulates innovation.

In terms of digital storage, Arm has always been committed to promoting the cutting-edge development of the storage industry, including improving performance, responsiveness, energy efficiency and security.

At CFMS2022, Arm will deliver a keynote titled "Arm: The Cornerstone of Future Cloud-Edge Digital Storage and Computing Innovation" where he will share his views and insights on industry trends, including how new technologies will transform the existing layout , and the capabilities of

Arm will partner to provide the best storage and storage solutions for computing.​

View attachment 15109

Registration for China Flash Market Summit CFMS2022 has been opened, and you can register for the conference through the summit official website (https://cfms.chinaflashmarket.com/) or "Flash Market" APP.

I had a little chuckle when I read what I assumed was a typo in “China Flash Memory Market Summit”. Or at least I hope it is a typo. Wouldn’t it be CFMMS rather than CFMS if that wording was correct? They seem to use both wordings (with and without the word memory) in their website.

Surely the world is aware of the limitations of Flash memory. Flash memory doesn’t seem to fit the Summit theme of “Explore the Unknown • Find the Future”.

RRAM is the future, and does fit that theme.

With the theme of "Explore the Unknown • Find the Future", they should be concentrating on alternatives to Flash memory. I see, and hope for, a near future that doesn’t include Flash Memory.

Flash memory will be to RRAM and MRAM what VCR is to DVD, or what floppy disc is to CD.

I understand this summit seems relevant because ARM is involved, especially considering Akida acceleration helps in the memory storage arena And with data centers. ARM just has its fingers in EVERY pie at the moment, even pies involving soon-to-be-dinosaurs like Flash memory.

Maybe this post would be more relevant in a WeebitNano forum, exalting how the world is still wasting time on this almost already outdated technology.

 

[Fullmoonfever]

Article on Top 10 Tech Trends by McKinsey & Co late June reprinted on the World Economic Forum site.

Won't dump all in the post but worth a look.

Personally think Akida has potential for direct impact into the top 8 and supporting impacts to 9 & 10.

Bodes very well imo.

Here are the top 10 trends in tech

These 10 tech developments are attracting the most venture capital and producing the most patent filings. They include AI algorithms that train machines.

www.weforum.org

 

[Slymeat]

Article on Top 10 Tech Trends by McKinsey & Co late June reprinted on the World Economic Forum site.

Won't dump all in the post but worth a look.

Personally think Akida has potential for direct impact into the top 8 and supporting impacts to 9 & 10.

Bodes very well imo.

Here are the top 10 trends in tech

These 10 tech developments are attracting the most venture capital and producing the most patent filings. They include AI algorithms that train machines.

www.weforum.org

View attachment 15123 View attachment 15124

I completely agree @Fullmoonfever—AI, the edge, and even neuromorphic computing proliferate all those trends.

And I learnt a new concept “QuantumAI“. That sounds real interesting. Yay I have a new concept to understand and in which to look for companies doing bleeding edge research. I expect Archer Materials may be one such company; the CEO Mohammad Choucair is a very clever fella, and runs a highly progressive company. Not to mention they already have a quantum chip!

It’s a pity that a trading scam using the name already exists.

I do like QuAIL though—Quantum Artificial Intelligence Lab—and the idea of using quantum computing to assist machine learning. I expect that might be the extent of it.

Will there ever be a CNN or even a SNN of quantum processors. In a way, that is sort-of how quantum computing works—every conceivable answer to a problem is computed in parallel and then a voting or weighting system declares the most appropriate answer. You are not guaranteed to get the same answer each time, but the answer you do get will still be correct!

And we all thought WANCAs were having a hard time understanding Akida.

 

[Fullmoonfever]

I completely agree @Fullmoonfever—AI, the edge, and even neuromorphic computing proliferate all those trends.

And I learnt a new concept “QuantumAI“. That sounds real interesting. Yay I have a new concept to understand and in which to look for companies doing bleeding edge research. I expect Archer Materials may be one such company; the CEO Mohammad Choucair is a very clever fella, and runs a highly progressive company. Not to mention they already have a quantum chip!

It’s a pity that a trading scam using the name already exists.

I do like QuAIL though—Quantum Artificial Intelligence Lab—and the idea of using quantum computing to assist machine learning. I expect that might be the extent of it.

Will there ever be a CNN or even a SNN of quantum processors. In a way, that is sort-of how quantum computing works—every conceivable answer to a problem is computed in parallel and then a voting or weighting system declares the most appropriate answer. You are not guaranteed to get the same answer each time, but the answer you do get will still be correct!

And we all thought WANCAs were having a hard time understanding Akida.

@Slymeat

Then you might like this read.

It popped up when I was doing some searching (Nintendo links haha) couple nights ago and doesn't involve Akida unfortunately so I just read it but have dug it up again for you.

Has brief bit on quantum but it's what the AI did that's interesting.

Gamingsym

Level UP

gamingsym.in

An artificial intelligence has discovered alternative laws of physics​

August 9, 2022 Aditya Khavanekar


A snip of couple things the researchers said....

Their work is based on a camera that observes the evolution of a physical system, such as a pendulum. And this is the one and only resource at his disposal. From these simplistic visual examples, the AI is responsible for determining the number of parameters needed to describe the behavior of the system in question. In a very colorful way, it’s a bit like a genius scientist rediscovering physics in real time in a parallel dimension.

The researchers were therefore curious to see if the AI would also find four parameters, which could possibly indicate that it would have followed the same reasoning as humans. But the answer proposed was very surprising: to describe the double pendulum, the system estimated that it would be necessary… 4.7 settings.

The researchers therefore repeated the experiment with other already well-documented mechanical systems. And each time, the outcome was the same: the algorithm consistently succeeded in predicting the evolution of the mechanical system based on entirely new variables that did not correspond to any parameter of Newtonian physics.

“ Without any prior knowledge of the physical mechanisms involved, our algorithm discovered the intrinsic dimensions of the observed dynamics and identified sets of state variables “, explain the researchers. In short, this AI doesn’t just think outside the box; she downright imagines new ways of getting around.

This was another write up of the same experiment and has some of the images / footage used to test plus the research doc link.

Hidden Variables

Force, Mass, Acceleration. These are the three variables that make up Newton’s iconic equation F=MA. But how did Newton know about these concepts in the first place? A precursor step to...

www.creativemachineslab.com