BRN Discussion Ongoing

D

Deleted member 118

Guest
BRN have done everything right to commercialize Akida.

Hired people from the big players, provided development kits, built an ecosystem etc.

Market timing is also spot on with early adoption commencing now & mass adoption expected within 3-6 years.

H2 FY23 is the tip of the iceberg for BRN. An Nvidia or Qualcomm IP licence agreement will spice up the SP.

View attachment 32648
An old picture from a while back, which has change a little bit, but still more on top of the iceberg than below

7E866CF1-B883-472A-A672-A82FDA0C75B6.jpeg
 
  • Like
  • Fire
  • Love
Reactions: 24 users
D

Deleted member 118

Guest
With past week’s avalanche of good news and dot-joining having perceptibly brightened the forum mood despite the continuous downward spiralling of the BRN share price, my rambling musings below - that I started jotting down last weekend - may seem somewhat obsolete and not particularly relevant to those among you that see Brainchip as a no-brainer stock anyway, since I would be preaching to the converted. Nevertheless, I would still like to invite you all to come with me on an imaginary journey into a lush green bamboo grove. Please excuse the unintended novella length and of course feel free to skip my post altogether.

One of my favourite childhood books was Leo Lionni‘s charming classic “Frederick“. In case you’ve never heard of it or your memory needs some refreshing, here is a short summary of this 1967 timeless picture book:

“Frederick is the story of a daydreamy field mouse who spends his summer days very differently from his mouse family. As the other mice work busily night and day to gather nuts, wheat, and straw, Frederick instead gathers sun rays, colors, and words. Later, when all of the supplies have been exhausted, and “corn was only a memory,” the mice ask Frederick for his contribution to the group. “Close your eyes,” Frederick tells them, climbing atop a tall rock. And from his stone perch he envelops the other mice in lush words that bring the warmth of the sun’s golden glow, that paint within their gray minds the bright colors of blue periwinkles, golden wheat, and berry bushes.”
(Kirsten Hall, children’s book agent and author).

To be honest, I very much feel like empty-handed Frederick here in this forum, where the vast majority of posters are heaps tech-savvier than me, working busily night and day to gather and share info on this undervalued gem of a company, which by many is predicted to be at the forefront of the 4th Industrial Revolution and is currently in the process of making itself indispensable to its ever increasing circle of customers and partners, true to its ambition to make edge AI ubiquitous and Akida its standard.

Prior to registering with TSE a couple of weeks ago, I had actually been a silent reader of this forum since its inception aka The Great Migration, but never really felt I could contribute anything of substance (other than accumulating shares, that is). Well, I guess now that we are currently stuck in the bleak midwinter of BRN‘s stock price performance, it is time for some (alas not as masterful as Frederick’s, I am afraid) story-telling contribution from my side then, to lift weary spirits. Some of you still have plenty of energy to withstand the freezing temperatures of BRN’s and the general stock market’s seemingly endless winter; at the same time others feel dejected, are shivering and afraid of starving, and while I can‘t offer you any nourishing wheat grains or tasty nuts, I will try to conjure up an imagery that will hopefully let you forget about your growling stomachs for a while, make you feel warm and fuzzy inside and help you visualise the sumptuous feast we are all drooling over. (Wait - won’t that lead to even more borborygmi aka tummy rumbles? 🤔 Oh, I suppose it is a vicious circle indeed; metaphors can be tricky.) Winter shall pass. Hopefully very soon, we will smell that spring is in the air.

Other posters before me have already likened the growth of a successful company to the agricultural concepts of planting, tending to and harvesting, but with most plants/trees/crops you will be able to observe fast and continuous growth - within days, weeks or months. The path of evolution taken by disruptive tech companies, on the other hand, is slightly different. I was reminded of this when I recently happened to come across the following video called ‘The Miracle of the Chinese Bamboo’ - I, for my part, found it worthwhile watching and hope you will enjoy it, too:



The word ‘miracle’ here is actually somewhat misleading, though, as it only APPEARS to be a miracle in the eye of the beholder who is not familiar with this plant‘s alleged botanic anomaly: no visible growth above the surface for more than five years despite being watered and tended to daily, followed by sudden explosive growth, reaching a height of almost 30 m within six weeks. While bamboo is indeed known for being the world’s fastest-growing plant, with some species growing just shy of a meter a day or so, the tale‘s extraordinary bamboo species doesn‘t exist in real life to the best of my knowledge. But it still makes for a powerful inspirational story that motivational speakers like to use as a parable about personal growth and success in life, illustrating the importance of steady and diligent work, perseverance and patience, even if evidence of progress is not immediately visible. An almost dated concept in our age of instant gratification, it seems.

To me the parable also constitutes an apt metaphor for the share price development of a disruptive technology super crop like Brainchip, except that it is actually not us shareholders doing the daily watering of the ‘spots’ in this huge field planted with millions of Chinese Bamboo seeds (> BRN shares), but instead by investing, we conveniently delegate this tedious task to professional gardeners (> the hard working team at Brainchip).

Below the surface, the network of roots & rhizomes has been growing for years and continues to do so day by day, establishing a solid foundation for its ultimate height and weight (> future rapid growth), a strong support structure that will eventually enable the plant to withstand even harsh conditions, although there is nothing yet visible overground (> the current share price not at all reflecting the company‘s actual progress and paradigm-shifting potential due to various reasons, including revolutionary tech yet to be fully understood and adapted, long product development timelines, NDAs in place, massive global economic headwinds as a result of the pandemic and Russia‘s invasion of Ukraine to name but two major underlying causes, relentless shorting attacks, panic selling by short-sighted retail shareholders dissatisfied with the lumpy revenue so far and and lamenting the lack of further signed deals, not enough buying interest from prospective shareholders for the same reasons and on top of that genuine shareholders forced to sell because they need money in this difficult economic climate/for unplanned personal reasons, and yet again others selling in order to get back in at a lower price, which could possibly backfire should a sudden price-sensitive announcement see the share price soar...).

Unlike the farmer in the parable, though, we shareholders - and in fact non-shareholders alike - have the fantastic option of monitoring parts of the expanding underground root network by means of a subterranean camera (Akida Inside, of course! No blurry pictures here, but razorsharp images…) to help us observe what is going on in the soil (> thanks to company press releases, ASX anns, quarterlies/half yearly reports/AGMs, podcasts, testimonials from other companies, including validation from big industry players, hiring of high calibre staff, articles and online likes & comments from tech experts worldwide, exhibition demos and last but not least the wealth of info gathered and generously shared in this amazing forum, which is extremely helpful for tech-challenged shareholders like me who don‘t have a semiconductor industry background or similar). While not a crystal ball, this subterranean camera ought to serve as objective reassurance to us that our seed is viable and of excellent quality. An advantage the farmer in our story did not have - he had to simply believe in the viability of his seeds and in the meaningfulness of faithfully watering the soil daily without proof of growth for five long years, during which he had to overcome mockery by his fellow villagers (> WANCAs) as well as many moments of self-doubt.

So we as BRN shareholders today are in a much better and derisked position! There is nothing much we can do for the time being (IMO) other than trusting the Brainchip team to continue to do their job well, be excited about the ever expanding ecosystem, weather the bear market storm and persevere in the face of the tumbling BRN share price (as long as your personal situation doesn‘t force you to sell any shares right now, which would be unfortunate indeed) as well as seize the window of opportunity to top up at these bargain prices, if your funds and personal diversification risk profile allow it. And then one day, in the not too distant future (bar a catastrophic black swan or black elephant event happening), instead of overlooking barren, sun-torched ground (> the share price looking right now), we will be standing in a lush green forest of giant bamboo stalks, providing us shade. Except that our very own miracle bamboo won’t stop growing at a height of 90 feet (almost 28 m), but can actually keep growing and growing (> think science fiction, disruptive technology, years ahead of competition, secret sauce, ubiquitous, IP moat, competitors turning into partners, future Akida generations, stream of royalties - you name it…) And not only is it highly likely that our investment will one day bear abundant fruit financially, but - equally significant to me - the low-cost, energy-efficient and high-performance Akida processor family will also be immensely beneficial to mankind on its way towards a more sustainable and greener future. So what is not to like - except for the current share price, that is? Close your eyes, imagine those towering green bamboo stalks gently swaying in the wind and don‘t forget to breathe…. (They are firm yet flexible, by the way, and won’t break, but rather bend to ground level in a storm and straighten back up afterwards. This truly is a botanical fact about bamboo, and also that so-called running bamboo will become ubiquitous in your garden if you let it grow unchecked.)

I suppose the Doubting Thomases among us will remain unconvinced until the bamboo shoots finally brake the surface and start growing to a towering height (> the share price soaring on that often quoted hockey stick curve, once the financials will show a substantial increase in revenue). But maybe some of you, while understandably frustrated with the current share price at a 52 week low (which to others like me actually feels like a gift from heaven, while we are happily topping up and in my case averaging down), will find this inspirational story of the Chinese Bamboo somewhat helpful in visualising Brainchip‘s projected exponential share price growth curve, although in essence, all of this has been said many times before by other contributors to this forum. Yet, as an avid photographer I know that it is often worthwhile to view things from a slightly different perspective. And while there will always be share price fluctuation on a day-to-day basis, I am confident the long-term upward trend will soon become blazingly obvious when zooming out a little, looking at the big picture.

And to those of you who dismiss this parable as just another feel-good story: perhaps it will still resonate with you as a general reminder of the importance of a solid foundation, steady work and patience in life to reap future rewards, even if there is not always instant evidence that your efforts are having any effect. They may not pay off every time, but more often than not they do. Sadly, this equally holds true for prophets of doom and their constant downramping.
Keep in mind, though, that the darkest hour is always before dawn.

FF, I hope you don’t mind me stealing your signature sign-off for today:

My opinion only DYOR

Frangipani

P.S.: Serendipitously, I discovered that Leo Lionni also wrote a book for adults called ‘Parallel Botany’, “a field guide to imaginary plants, which Lionni presents with the authority of an academic writer - peppering his writing with references to real places and people, just to complicate things“ (Daisy Alioto: ‘Parallel Botany‘ in the Age of Alternative Facts). Now doesn‘t that sound he would have just loved the miraculous Chinese Bamboo?

Probably a great read, but far too much for my little brain to take in, but going on what everyone has written about it I’ll give it a big

 
  • Haha
  • Like
  • Love
Reactions: 14 users

mrgds

Regular
Probably a great read, but far too much for my little brain to take in, but going on what everyone has written about it I’ll give it a big




You dont have to read the post R577 ................... the video is the essence of the written word.
Even a WHU fan should be able to understand the video ...................:rolleyes: .......................😅

AKIDA BALLISTA
 
  • Haha
  • Like
Reactions: 5 users

Deadpool

Did someone say KFC
A lovely post


Thank you Frangipani for your lovely symbolism and beautifully written contribution.
Another example of the strength and resilience of our group as expressed by its diversity, allowing for the cross pollination of contacts, ideas and concepts.
In these days of growing specialisation it is getting difficult to embrace, adopt and maintain a wider view.
Indeed our growing acceptance and reliance on social media is channeling more and more of us down predetermined, algorithmically defined paths that seperate us and challenge long held social mores and societal shared truths and visions.
So, it is nice to belong to a group with a shared interest and common advocacy and be able to rejoice in its community.

So thanks to all the techies, technophobes, geeks, greeks, cat people, neuromorphomaniacs, joiners of dots, calculators of imaginary figures, ogres, professional amateurs, sleuths, jokers, gif and meme painters and the rest of you ragtag contributors and those who just like to watch.

Look forward to meeting with some of you again and some for the first time at this years AGM.

It is a really wonderful opportunity to look our Directors and Management in the eye, shake their hands, hear what they have to say,
and thank them.
I like to watch.;) I'm not weird and creepy at all, I'm cool and awesome.:giggle:

cat spying GIF
 
  • Haha
  • Like
  • Wow
Reactions: 16 users
D

Deleted member 118

Guest
You dont have to read the post R577 ................... the video is the essence of the written word.


AKIDA BALLISTA
What Video



COYI
 
  • Haha
  • Like
Reactions: 5 users

mrgds

Regular
What Video



COYI

 
  • Like
Reactions: 1 users
D

Deleted member 118

Guest
plus i probably missed it as it was somewhere in the middle of all that writing
 
  • Haha
  • Sad
  • Like
Reactions: 4 users

mrgds

Regular
plus i probably missed it as it was somewhere in the middle of all that writing

:rolleyes: ..................................... im just thinking/hoping @Frangipani is one of our BOD/CT/F/M O"s giving us shareholders of around 5yrs the heads up
on the value of patience. :cool: LOL

Hope this helps/works, ............. just make the effort to click on the big red circle in the middle, ....................:p

 
  • Like
  • Haha
  • Fire
Reactions: 7 users

Steve10

Regular
At 3min 54secs in video the Texas Instruments rep mentions that they have added a TI developed AI engine inside their new vision processors. Based on ARM tech with low power, low cost, low latency & edge mentioned.

It's their new AM62A (2 TOPS), AM68A (8 TOPS) & AM69A (32 TOPS) vison processors with AI.

Underneath the AM series chip demos are their partners projects using the tech.

Can identify iGlobal Solutions, Avnet, ieasy, Variscite, Toradex, Open Standard Module, Forlinx, Critical Link, Octavo Systems, Congatec, Phytec, SolidRun, imagimob.

At 9min 36 secs he shows the retail application with object classification for fast checkout.

 
  • Like
  • Love
  • Fire
Reactions: 33 users
D

Deleted member 118

Guest
:rolleyes: ..................................... im just thinking/hoping @Frangipani is one of our BOD/CT/F/M O"s giving us shareholders of around 5yrs the heads up
on the value of patience. :cool: LOL

Hope this helps/works, ............. just make the effort to click on the big red circle in the middle, ....................:p


I did press the big red button, that was the problem.

 
  • Haha
  • Like
  • Sad
Reactions: 11 users

TECH

Regular
Good morning/afternoon/evening, wherever you are viewing this platform.

Is this current window of opportunity rapidly coming to an end, well, I believe so, 45c is a complete fabrication of the truth, things are about
to explode on a worldwide scale (my educated opinion).

Tech Ecosystem


This is the biggest change of direction that our company embarked on within months of Sean taking over the reigns, he
is passionate, competitive, well researched, is etching out a direct path to success for the benefit of all.

Many would have seen a number of years ago the slide of a potential roadmap unveiling the future direction of AKIDA,
that is, AKD 1 through to AKD 10, being (AGI).

Last week I spoke with Peter briefly to ask him to please explain or more to the point, clarify what he actually meant when
referring to AGI, as many would already know, the Singularity suggests a future world where humans lose their sense of
reality and technology gets out of control, Peter was very clear in expressing his views on this matter, with the emphasis on
Beneficial AI .

"We don't expect to achieve AGI by 2030"

There may be more than 7 chips taped out, there maybe IP's developed for specific customers, I really get the sense that we are
being extremely flexible and accommodating, yes we are a young start-up, and it's natural to want to please all potential clients,
so this business attitude will be respected to the point that we are going to end up in a much stronger, commanding position
in the overall market.

After receiving Peter's reply I had to readdress my own understanding, without realizing it as such, I was still locked into Lou's
earlier roadmap, which has changed for the better in my personal opinion.

Another area where my own understanding was wrong, was in the fact that, despite us changing direction rather sharply 3 months prior
to the last AGM, we are still potentially going to be supplying NSoC for specific customers for specific end user cases (products).

Finally, may I say that @Steve10 you have become a very strong contributor on this forum, you present excellent material and are
nicely balanced in your views, your figures give many food for thought, so from me personally, I say thank you. (y)

From a rather hot day in Perth, I say cheers everyone. Tech :cool:
 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 79 users
Amazing how only 3 years ago you could not find anything about brainchip on the web.

Now we are everywhere.

Our time has come BRN team!
Yes there is some serious media momentum building now from many channels and partner companies. Very exciting to see.

It is hard not to start get further excited around the many potential revenue streams and also what could be in the next quarterly.

We know we were told to watch the revenue late last year ... it didn't really come through like we expected - is the larger revenue flow just late and still expected .....this quarterly?

If I were the CEO/CIO and heading into the next AGM in two months I know what I would be doing - the order of the day now is securing dollars, deals or clearing NDA covenants/clauses in order to show something decent in the next quarterly report or an ASX market release or two to show commitment of larger partners and larger dollars to flow so that the company can start to be effectively valued on forward earnings.

I am honestly expecting to see something from Sean and team to allow the company's real value to start to be reflected in the SP otherwise I think they know the AGM is not going to be a fun exercise.

There is also a heck of a lot of ICT Professional Services (Staff Labour) being expended here for many many partners and technology companies for many years in some instances - are BRN going to keep offering their ICT Professional Services for free to help develop these third party products - I would expect to see a few million dollars or more collected here and there as a straight up labour cost recovery and also when certain product development milestones are achieved.

Basic revenue sources I see are:
  1. IP Licences
  2. Hardware/Chips (not a lot here really)
  3. IP Royalties
  4. Engaged ICT Professional Services (Labour)
  5. Development and Delivery Milestones for Partner (combo of IP and Labour)
What is going to drop out this quarter? Only 4 to 5 weeks to go before we know.

On another specific note: We know the new MBOS with "Akida inside" is set to be in the 2024 Mercedes Benz range - these 2024 cars are usually released to the market for starting purchase consumption in the year before in lead up - which should be around Sept/Oct 2023 typically. You would have to think that MB are starting to actually produce the components now in order to have complete cars driving out the door in October this year for purchase(6.5 months time). We know car companies have parts physically built at various factories all around the world then shipped to one end factory for the end car build. Is there any MB revenue starting up in this space this quarter?

Development milestones payments could be forthcoming from Mercedes to their partners with flow of royalties starting due to these physical products being built .....the flow of dollars has to be close from Mercedes specifically.

Thoughts?
 
Last edited:
  • Like
  • Fire
  • Love
Reactions: 37 users

Steve10

Regular
At 3min 54secs in video the Texas Instruments rep mentions that they have added a TI developed AI engine inside their new vision processors. Based on ARM tech with low power, low cost, low latency & edge mentioned.

It's their new AM62A (2 TOPS), AM68A (8 TOPS) & AM69A (32 TOPS) vison processors with AI.

Underneath the AM series chip demos are their partners projects using the tech.

Can identify iGlobal Solutions, Avnet, ieasy, Variscite, Toradex, Open Standard Module, Forlinx, Critical Link, Octavo Systems, Congatec, Phytec, SolidRun, imagimob.

At 9min 36 secs he shows the retail application with object classification for fast checkout.



Have a look at pricing for TI's AM series vision processors with AI.

Top of range AM69a with pricing in USD.

XAM69A98ATNGHAALY​

ACTIVE

32 TOPS vision SoC for 1-12 cameras, Autonomous Mobile Robots, Machine Vision, Mobile DVR, AI-BOX​

Pricing​

QtyPrice (USD)
1 — 99$174.000
100 — 249$157.500
250 — 999$153.000
1,000 +$150.000


Mid range AM68A are USD $29.90 to $49.63

Low range AM62A are USD $18.74 to $29.09
 
  • Like
  • Love
  • Wow
Reactions: 28 users

Terroni2105

Founding Member
A Spanish guy unboxing Akida on Youtube.



He is an Edge Impulse platform user.



He's got 32,000 subscribers to his YouTube channel, great exposure (y)
 
  • Like
  • Love
  • Fire
Reactions: 38 users

White Horse

Regular
I posted this research paper some time ago but the KEY PARTNERSHIP with PROPHESEE makes it more relevant:


My opinion only DYOR
FF

AKIDA BALLISTA
Hi FF.
On reading this the second time around, I have an idea that they may be referring to Akida Ver 2.0. Note the use of 8 bit.

A. TARGET HARDWARE
StereoSpike has resolutely been developed in the philosophy
of spiking neural networks. As a result, it is essentially
implementable on dedicated neuromorphic hardware, such as
Intel Loihi [7], IBM TrueNorth [8]. These chips can leverage
the binarity and sparsity of spike tensors navigating through
the network. In addition, we believe that our model being
feedforward and requiring a reset on all of its neurons at each
timestep is not a problem, because resetting membrane poten-
tials is actually less costly than applying a leak. Therefore,
statelessness can be seen as an advantage over recurrence in
spiking models with similar performances. However, we are
aware that current neuromorphic chips are initially designed
for the implementation of stateful units, and acknowledge that
we do not leverage this feature. Consequently, we believe
that it rather fits to dedicated hardware for stateless models
with sparse quantized activations. We therefore consider that
Brainchip’s Akida chip [9] is a good fit. As it imposes
weights to take at most 8 bit, we quantized StereoSpike’s
weights using PyTorch natively available post-training static
quantization. The process resulted in an even lighter model
with 8 bit wide unsigned integer weights, for the price
of a minor performance drop (i.e., MDE of 17.1 cm
on indoorflying split 1). Presumably, quantization-aware
training would do even better. This demonstrates the efficient
deployability on such hardware. Finally, we would like
to emphasize that our class of model with sparse binary
activations and less constrained weights provides a good
compromise between Spiking Neural Networks (SNNs) and
Binary Neural Networks (BNNs).
 
  • Like
  • Fire
Reactions: 29 users

Violin1

Regular
Good morning/afternoon/evening, wherever you are viewing this platform.

Is this current window of opportunity rapidly coming to an end, well, I believe so, 45c is a complete fabrication of the truth, things are about
to explode on a worldwide scale (my educated opinion).

Tech Ecosystem


This is the biggest change of direction that our company embarked on within months of Sean taking over the reigns, he
is passionate, competitive, well researched, is etching out a direct path to success for the benefit of all.

Many would have seen a number of years ago the slide of a potential roadmap unveiling the future direction of AKIDA,
that is, AKD 1 through to AKD 10, being (AGI).

Last week I spoke with Peter briefly to ask him to please explain or more to the point, clarify what he actually meant when
referring to AGI, as many would already know, the Singularity suggests a future world where humans loose their sense of
reality and technology gets out of control, Peter was very clear in expressing his views on this matter, with the emphasis on
Beneficial AI .

"We don't expect to achieve AGI by 2030"

There may be more than 7 chips taped out, there maybe IP's developed for specific customers, I really get the sense that we are
being extremely flexible and accommodating, yes we are a young start-up, and it's natural to want to please all potential clients,
so this business attitude will be respected to the point that we are going to end up in a much stronger, commanding position
in the overall market.

After receiving Peter's reply I had to readdress my own understanding, without realizing it as such, I was still locked into Lou's
earlier roadmap, which has changed for the better in my personal opinion.

Another area where my own understanding was wrong, was in the fact that, despite us changing direction rather sharply 3 months prior
to the last AGM, we are still potentially going to be supplying NSoC for specific customers for specific end user cases (products).

Finally, may I say that @Steve10 you have become a very strong contributor on this forum, you present excellent material and are
nicely balanced in your views, your figures give many food for thought, so from me personally, I say thank you. (y)

From a rather hot day in Perth, I say cheers everyone. Tech :cool:
Nice post @TECH - the potential for chip provision is really interesting as I thought that was completely out the window. Btw - I love how you've named the central ecosystem after yourself......lol
 
  • Haha
  • Like
Reactions: 10 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
Anyone up for knitting an AKIDA powered jumper for me for Christmas.

The following paper is another of those industries that do not yet exist AKIDA powered textiles:

“V. NEUROMORPHIC COMPUTING PROPERTIES
Neuromorphic computing concept originated in the 1980’s. Taking inspiration from computer science, mathematics to bio-inspired models of neural network. This emerging inter- disciplinary research field has the potential to disrupt tradi- tional computing methods and architectural implementations leading to a more centralized and combined memory and computational driven approach, moving away from the von Neumann architectural approach with separate memory and computing capabilities and high compute power needs. Such inspiration coming from the working of the human brain paves the way for new and more fault tolerant layered and parallel architectural designs and layouts.
FIGURE 7. Timeline of technological advancements in Neuromorphic Computing
In order to understand what aspects of neuromorphic computing can inspire innovative advancements in the next generation of smart etextiles and on-garment edge based intelligence, we will first highlight the core key architectural elements of importance within neuromorphic computing and assess their potential within an etextiles domain. Neuromor- phic computing core architecture is based on the concept of communicating through event driven spikes generated through simple processing structures represented by synapses and neurons. Ongoing research is pushing the production possibilities using complementary metal oxide semiconduc- tor (CMOS) technology to develop neuromorphic spiking neural network hardware implementations [45][46][47]. Key properties such as size, weight, low power consumption, and modular design (scalability) are dominating the research areas of focus linked to such technologies. Over the years advancements in CMOS technology has driven smaller and more power efficient systems with the capability to mass produce. Such technology combined with advanced machine learning techniques has directly lead to the simulation and implementation of silicon based neurons, otherwise defined as neuromorphic computing.
[Fig 7] highlights advancements, with Brainchip (https://brainchip.com/) announcement in 2022 claiming to be the worlds first commercial producer of a Neuromorphic AI processor ‘Akida’ that has the capability to mimic the working of the human brain and process data with high precision and energy efficiency. Akida being an event-based AI neural processor featuring 1.2 million neurons and 10 billion synapse.”


My opinion only DYOR
FF

AKIDA BALLISTA

PS: Navy blue, crew neck, Cable knitt to go with jeans. Thanks.

Sure @FactFinder! I don't know how to knit but I'm always up for a bit of a challenge! What size are you? I mean what are your dimensions? I'll have to get started soon if I want to get this to you by Christmas.


Screen Shot 2023-03-20 at 4.31.5.png
 
  • Haha
  • Like
  • Love
Reactions: 17 users
Hi FF.
On reading this the second time around, I have an idea that they may be referring to Akida Ver 2.0. Note the use of 8 bit.

A. TARGET HARDWARE
StereoSpike has resolutely been developed in the philosophy
of spiking neural networks. As a result, it is essentially
implementable on dedicated neuromorphic hardware, such as
Intel Loihi [7], IBM TrueNorth [8]. These chips can leverage
the binarity and sparsity of spike tensors navigating through
the network. In addition, we believe that our model being
feedforward and requiring a reset on all of its neurons at each
timestep is not a problem, because resetting membrane poten-
tials is actually less costly than applying a leak. Therefore,
statelessness can be seen as an advantage over recurrence in
spiking models with similar performances. However, we are
aware that current neuromorphic chips are initially designed
for the implementation of stateful units, and acknowledge that
we do not leverage this feature. Consequently, we believe
that it rather fits to dedicated hardware for stateless models
with sparse quantized activations. We therefore consider that
Brainchip’s Akida chip [9] is a good fit. As it imposes
weights to take at most 8 bit, we quantized StereoSpike’s
weights using PyTorch natively available post-training static
quantization. The process resulted in an even lighter model
with 8 bit wide unsigned integer weights, for the price
of a minor performance drop (i.e., MDE of 17.1 cm
on indoorflying split 1). Presumably, quantization-aware
training would do even better. This demonstrates the efficient
deployability on such hardware. Finally, we would like
to emphasize that our class of model with sparse binary
activations and less constrained weights provides a good
compromise between Spiking Neural Networks (SNNs) and
Binary Neural Networks (BNNs).
Absolutely.

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Love
  • Haha
Reactions: 10 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
  • Like
  • Thinking
  • Love
Reactions: 8 users

Murphy

Life is not a dress rehearsal!
Sure @FactFinder! I don't know how to knit but I'm always up for a bit of a challenge! What size are you? I mean what are your dimensions? I'll have to get started soon if I want to get this to you by Christmas.


View attachment 32657

It's not that easy to pull the wool over FF's eyes, Miss Kitty!! :) reminds me of the 1970's graffiti where someone wrote on the walls of the university toilets "My mother made me a homosexual" and someone added underneath "If I send her the wool, will she make me one?" :)


If you don't have dreams, you can't have dreams come true!
 
  • Haha
  • Like
Reactions: 11 users
Top Bottom