BRN Discussion Ongoing

VictorG

Member
My wife said that if I don't stop watching the stock market and mow the grass she'll slam my head on the keyboard, but I think she's jokinfjreoiwjrtwe4to8rkljreun8f4ny84c8y4t58lym4wthylmhawt4mylt4amlathnatyn
 
  • Haha
  • Like
  • Love
Reactions: 60 users
We have had the original report from Sandia National Laboratories but this further reporting reminded me of the former CEO Mr. Dinardo stating that Peter van der Made had flown to Europe to meet with representatives of a company in the financial industry to discuss what AKIDA technology had to offer.

I particularly noted that having more neurons on a chip was the margin of victory to expanding the use of SNN networks. Hello AKD1000 at 1.2 million neuron equivalents and 10 billion synapse and currently capable of connecting 64 as configured it is optimally situated to assist a company like DELL Technologies be first in this space and designed to allow for up to 1024 if anybody ever wanted to take over the world. Just a heads up Puto next time you know who to call.

I just had a flash back to a troll on the other place (cannot remember it’s name probably because I have no interest) who claimed to be in the semiconductor industry and argued with and ridiculed anyone who suggested that AKIDA technology could be used in data centres and the cloud.

Clearly this thing was deliberately spreading false information in an attempt to undermine posters such as @uiux.

Must once again thank @zeeb0t for creating this safe harbour.

My opinion only DYOR
FF

AKIDA BALLISTA

 
  • Like
  • Love
  • Fire
Reactions: 49 users

miaeffect

Oat latte lover
My wife said that if I don't stop watching the stock market and mow the grass she'll slam my head on the keyboard, but I think she's jokinfjreoiwjrtwe4to8rkljreun8f4ny84c8y4t58lym4wthylmhawt4mylt4amlathnatyn
romper-dance.gif

Get this guy to do your lawn. Wife will let you stay in your room and do whatever you want.
 
  • Like
  • Haha
  • Love
Reactions: 23 users

Diogenese

Top 20
This is the link to the above paper:


It might be useful to newer investors to understand the Brainchip roadmap as far as technology advancement is concerned.

AKD1000 the first in the series is despite running on virtually no power and being the only standalone chip in the world capable of one shot, few shot and incremental learning as well as being able to process all five senses as well as radar, Lidar and Ultrasound is likely to be seen as unimpressive when AKD2000 incorporating LSTM is released most likely end of this year.

Why will AKD2000 be seen in this light? Well that is because what it will be capable of has been benchmarked by none other than Elon Musk. In an interview Elon Musk spoke to the problem of having compute capable of extrapolating that a plastic bag blowing across the road ahead does not require emergency braking or avoidance action. The plastic bag is a symbol of the present problem.

The ability of compute to extrapolate and decide a vast range of similar problems like the plastic bag or a cricket ball rolling onto the roadway from between parked cars means it could be followed by a child compared with a cricket ball rolling onto the roadway in other circumstances is a holy grail of developers of autonomous robots such as AV’s.

Peter van der Made has spoken to the intention that AKD2000 with LSTM will resolve this compute issue.

Coincidentally, though some thought not, Peter van der Made actually spoke of the plastic bag problem being solved by AKD2000.

This massive achievement however will be dwarfed by the AKD3000/4000/5000 series whichever one it is that incorporates the cortical column.

In May last year Brainchip announced that Emeritus Professor Alan Harvey was joining the Brainchip Scientific Advisory Board. Professor Harvey has studied and published on the brain and cortical columns for many years.

His appointment came a little more than a month after Peter van der Made stated he and his team had a working cortical column on the bench at the Brainchip Perth Research Centre.

When Brainchip announces the generation of AKIDA technology chip that is using their cortical column capable of all the current technology features plus ‘perception, memory, expectation or decision making’ it will be a day that will be marked down in history as the start of a new technology age.

Rob Telson is correct to say Brainchip is only just getting started, we are at the tip of the iceberg and these are exciting times.

Brainchip is on a trajectory to success that if achieved will be unlike anything ever seen before in the commercial and scientific world.

My opinion only DYOR
FF

AKIDA BALLISTA
Hi FF,
This paragraph at the bottom of column 1 of page 15 discusses memory.

The original motivation for this model was to obtain a realistic description of associative memory occurring on top of spontaneous brain activity with the latter being a global attractor of the dynamics. In a later version in which neural clusters can overlap (modeling the fact that real neurons can code for more than one stimulus), this model is the closest biologically plausible analog so far of the influential Amari–Hopfield network (discussed below in Sec. VB3) and can store an extensive number of stimuli modeled as patterns of firing rates across neurons.

My neuroscientific layman's guess is that, in Akida, the model libraries and associated weights are equivalent to memory. Coding neurons for more than one stimulus is performed in Akida by selecting a different model from the library.

So my question is, can Akida be, to some extent, adapted to cortical neuron configuration by configuration of the data in the model, and to what extent is hardware modification required to enable cortical operation? Does this involve modification of the NPUs or the interconnexion matrix, or both?

The existence of this kind of metastable dynamics in a spiking network was first pointed out in Refs. 16 and 17 who noticed that, because of metastable activity, the network produces slow fluctuations in the neural activity, much slower than the time scales of the single neurons—the origin of such timescales is a long-standing problem in theoretical neuroscience.170–172 They also showed that, unlike the case of a homogeneous excitatory population, a stimulus will suppress trial-to-trial fluctuations, another widespread phenomenon in cortical circuits.173 Reference 22 also found that, in this clustered network, the dimensionality of the neural activity (Sec. VA3) is larger during ongoing metastable dynamics than when the network is externally stimulated.

I think what this is saying is that concentration suppresses idle thoughts, much like changing from an idling motor to engaging drive. Or, in the case of Akida, the model is loaded, but there is no input. There may be degrees of concentration, like thinking about doing something and actually doing something.

Reminds me of the time I was taking a young lady for a drive in a rather sporty automobile, and she mentioned that the engine was rather noisy. I explained that that was the acoustic tachometer ... you know that look you get ...
 
  • Like
  • Haha
  • Love
Reactions: 20 users

Slymeat

Move on, nothing to see.
I learnt a long time ago that academics proposing to provide a review of an entire area of science are having a lend and just trying to keep up their academic publication numbers which are apparently important in CV's for promotion in their world.

Poor old Peter van der Made would not get a job based on publications at the moment as he has been in stealth mode for years now.

I only posted the extract as being the easy read as I have searched for ages for someone prepared to encapsulate the outcome of understanding and creating a simulation of a cortical column. As for the maths I always assume that is correct and peer reviewed as I could not check it anyway.

The philosophy of whether we will ever understand the human brain using the human brain to derive an understanding and allowing for the variation in human brains and ongoing evolution is an intellectual quandary that I often mull over and cannot resolve. It does seem the more those engaged in this endeavour learn the less they know.

Personally I think that the human brain is a flawed model upon which to build useful artificial intelligence. The human brain makes mistakes constantly hence the saying "To err is human to forgive divine" or something like that. Being human not sure if I am misquoting but I will quote it nonetheless as I am using a flawed human brain to construct this response.

What use is an intelligent machine that perfectly mimics the human brain down to the making of inexplicable mistakes and the ego to keep going at it regardless.

In Autonomous EV's the push is in large part so that they can remove human drivers and eliminate accidents. Go figure that one out when you read about attempts to emulate the human brain to create truly autonomous vehicles. I am not Einstein but the logic breaks down in my opinion.

In short the artificial general intelligence that Peter van der Made is seeking must at some point divert from the position of fully replicating the human brain and as such a complete understanding of the human brain and how it works is unnecessary to this goal in my opinion but what would I know I am human after all and constantly right and wrong unless you ask my wife who will tell you I am constantly wrong unless I agree with her unless it comes to the law and investment and doing our tax planning.

My opinion only DYOR
FF

AKIDA BALLISTA
Well said @FactFinder. I’d be happy if the world stopped trying to emulate the way our brain works and rather concentrated on emulating what I believe to be a true definition of intelligence. The art of using past experience to arrive at a reasonable conclusion for a new problem, and learning from that experience.

Doing the above in isolation is even a better measure of intelligence.

So what IS NOT Artificial Intelligence:
- Looking up an answer from a huge lookup table/data set is not a measure of intelligence, and is not AI.
- Doing something fast is not a measure of intelligence, and is not AI.
- Doing something complicated is not necessarily intelligence nor AI.

I do like the plastic bag analogy, and I do remember that being spoken of before. Determining that a plastic bag floating in front of an EV is not a dramatic event in need of drastic behaviour to avoid, is bordering on a level of artificial intelligence that I would like to experience. It’s a nice target—for want of a better word!

Working the way a human brain works, as in neuromorphic processors, brings efficiencies to the table.
One shot learning, or even just the ability to learn on the fly, brings other efficiencies to the table.

These are areas in which Akida excels. And this DOES border on AI.

I particularly like the idea that solutions involving Akida only rely on seed programming. The limitations of a human programmer not being able to code for unknowns has been a concern of mine for a very long time. AI solutions need the ability to adapt to situations that were not known of when they were first conceived.
 
  • Like
  • Love
  • Fire
Reactions: 13 users

Diogenese

Top 20
I learnt a long time ago that academics proposing to provide a review of an entire area of science are having a lend and just trying to keep up their academic publication numbers which are apparently important in CV's for promotion in their world.

Poor old Peter van der Made would not get a job based on publications at the moment as he has been in stealth mode for years now.

I only posted the extract as being the easy read as I have searched for ages for someone prepared to encapsulate the outcome of understanding and creating a simulation of a cortical column. As for the maths I always assume that is correct and peer reviewed as I could not check it anyway.

The philosophy of whether we will ever understand the human brain using the human brain to derive an understanding and allowing for the variation in human brains and ongoing evolution is an intellectual quandary that I often mull over and cannot resolve. It does seem the more those engaged in this endeavour learn the less they know.

Personally I think that the human brain is a flawed model upon which to build useful artificial intelligence. The human brain makes mistakes constantly hence the saying "To err is human to forgive divine" or something like that. Being human not sure if I am misquoting but I will quote it nonetheless as I am using a flawed human brain to construct this response.

What use is an intelligent machine that perfectly mimics the human brain down to the making of inexplicable mistakes and the ego to keep going at it regardless.

In Autonomous EV's the push is in large part so that they can remove human drivers and eliminate accidents. Go figure that one out when you read about attempts to emulate the human brain to create truly autonomous vehicles. I am not Einstein but the logic breaks down in my opinion.

In short the artificial general intelligence that Peter van der Made is seeking must at some point divert from the position of fully replicating the human brain and as such a complete understanding of the human brain and how it works is unnecessary to this goal in my opinion but what would I know I am human after all and constantly right and wrong unless you ask my wife who will tell you I am constantly wrong unless I agree with her unless it comes to the law and investment and doing our tax planning.

My opinion only DYOR
FF

AKIDA BALLISTA
"To err is human ..."

So now you want the god chip?
 
  • Haha
  • Love
  • Like
Reactions: 7 users

Bravo

If ARM was an arm, BRN would be its biceps💪!
Here's another Valeo LinkedIn post from a week ago which was about the Start-Up & Innovation Summit.


Screen Shot 2022-03-31 at 12.55.43 pm.png


This is a translated portion from the interview with Geoffrey Bouquot (Valeo) in which he responds to the following question:

Is AI essential for you?

We have created a research laboratory, Valeo.ai, to make our own contribution in the field. It has around thirty researchers, and acts as an academic bridgehead for all our research in artificial intelligence. And we are part of the "trusted AI", encouraged by the European Commission and our government, to develop repeatable and explainable technologies, with major partners such as Thales, Orange... with a vision of European sovereignty.


The Trusted AI that Geoffrey is talking about was launched as part of the Confiance.ai program on 1 July 2021. Here's some interesting points about Trusted AI which can be found on the second link.

  • Thirteen industrial companies and academic partners have come together to form a collective that will meet the following major challenge: “Secure, make reliable and certify systems based on artificial intelligence”. They all want to industrialize and integrate trusted AI into their critical products and services while providing a technical framework for the future European AI regulation proposal.
  • Between 2021 and 2024, the Confiance.ai collective will focus on the design of a sovereign, open, interoperable and sustainable software tool platform enabling the integration of AI in critical products and services in a safe, reliable and secure way, which will be dedicated to the engineering of innovative industrial products and services integrating AI.
  • The first application sectors targeted by the initiative will be automotive, aeronautics, energy, digital, industry 4.0, defence and maritime, with various use cases such as online industrial control, autonomous mobility or decision support systems.
  • This is a collective of thirteen major French industrial and academic partners: Air Liquide, Airbus, Atos, Naval Group, Renault, Safran, Sopra Steria, Thales, Valeo, as well as the CEA, Inria, the IRT Saint Exupéry and the IRT SystemX.

I'd be quite surprised if the collective has not got around to discussing Akida yet...

 
  • Like
  • Fire
Reactions: 16 users
Hi FF,
This paragraph at the bottom of column 1 of page 15 discusses memory.

The original motivation for this model was to obtain a realistic description of associative memory occurring on top of spontaneous brain activity with the latter being a global attractor of the dynamics. In a later version in which neural clusters can overlap (modeling the fact that real neurons can code for more than one stimulus), this model is the closest biologically plausible analog so far of the influential Amari–Hopfield network (discussed below in Sec. VB3) and can store an extensive number of stimuli modeled as patterns of firing rates across neurons.

My neuroscientific layman's guess is that, in Akida, the model libraries and associated weights are equivalent to memory. Coding neurons for more than one stimulus is performed in Akida by selecting a different model from the library.

So my question is, can Akida be, to some extent, adapted to cortical neuron configuration by configuration of the data in the model, and to what extent is hardware modification required to enable cortical operation? Does this involve modification of the NPUs or the interconnexion matrix, or both?

The existence of this kind of metastable dynamics in a spiking network was first pointed out in Refs. 16 and 17 who noticed that, because of metastable activity, the network produces slow fluctuations in the neural activity, much slower than the time scales of the single neurons—the origin of such timescales is a long-standing problem in theoretical neuroscience.170–172 They also showed that, unlike the case of a homogeneous excitatory population, a stimulus will suppress trial-to-trial fluctuations, another widespread phenomenon in cortical circuits.173 Reference 22 also found that, in this clustered network, the dimensionality of the neural activity (Sec. VA3) is larger during ongoing metastable dynamics than when the network is externally stimulated.

I think what this is saying is that concentration suppresses idle thoughts, much like changing from an idling motor to engaging drive. Or, in the case of Akida, the model is loaded, but there is no input. There may be degrees of concentration, like thinking about doing something and actually doing something.

Reminds me of the time I was taking a young lady for a drive in a rather sporty automobile, and she mentioned that the engine was rather noisy. I explained that that was the acoustic tachometer ... you know that look you get ...
My immediate thought was the difference between the amount of concentration I give to your posts and the rocket memes of Rocket. LOL

I actually get what you are saying strangely enough.

One of the things I have wondered about is monitoring for noise and or vibration. Your sportscar analogy reminded me of it. Most of us have had the experience of someone new such as your young lady getting into our car and a couple of hundred metres down the road being asked what is that knocking sound or tap or something. Replying I can't hear anything out of the ordinary only to find later that some part had been wearing away over a long time and it was the cause of the noise detected by the young lady and not noticed by ourselves.

I have explained this away in the past to myself on the basis that it occurred so gradually over time that I became used to this noise. So the question I have is why would this not happen to an intelligent device monitoring for noise or vibration in a motor? The answer is I assume that there must be some way to set it to remember a perfect state and that any variation or noise outside that perfect state will be what it detects.

Now herein is my question. As no two mechanical devices will be in perfect harmonic balance one with the other it seems to me you would need to have an intelligent device that can learn that perfect state on the factory floor so to speak for the particular piece of machinery as it would be impossible to program every piece of individual piece of machineries noise or vibration perfect state in a central memory. In other words you ideally need on chip training preferably one or several shot in nature. Do you disagree?

My opinion only DYOR
FF

AKIDA BALLISTA
 
  • Like
  • Fire
Reactions: 15 users

Diogenese

Top 20
Hi FF,
This paragraph at the bottom of column 1 of page 15 discusses memory.

The original motivation for this model was to obtain a realistic description of associative memory occurring on top of spontaneous brain activity with the latter being a global attractor of the dynamics. In a later version in which neural clusters can overlap (modeling the fact that real neurons can code for more than one stimulus), this model is the closest biologically plausible analog so far of the influential Amari–Hopfield network (discussed below in Sec. VB3) and can store an extensive number of stimuli modeled as patterns of firing rates across neurons.

My neuroscientific layman's guess is that, in Akida, the model libraries and associated weights are equivalent to memory. Coding neurons for more than one stimulus is performed in Akida by selecting a different model from the library.

So my question is, can Akida be, to some extent, adapted to cortical neuron configuration by configuration of the data in the model, and to what extent is hardware modification required to enable cortical operation? Does this involve modification of the NPUs or the interconnexion matrix, or both?

The existence of this kind of metastable dynamics in a spiking network was first pointed out in Refs. 16 and 17 who noticed that, because of metastable activity, the network produces slow fluctuations in the neural activity, much slower than the time scales of the single neurons—the origin of such timescales is a long-standing problem in theoretical neuroscience.170–172 They also showed that, unlike the case of a homogeneous excitatory population, a stimulus will suppress trial-to-trial fluctuations, another widespread phenomenon in cortical circuits.173 Reference 22 also found that, in this clustered network, the dimensionality of the neural activity (Sec. VA3) is larger during ongoing metastable dynamics than when the network is externally stimulated.

I think what this is saying is that concentration suppresses idle thoughts, much like changing from an idling motor to engaging drive. Or, in the case of Akida, the model is loaded, but there is no input. There may be degrees of concentration, like thinking about doing something and actually doing something.

Reminds me of the time I was taking a young lady for a drive in a rather sporty automobile, and she mentioned that the engine was rather noisy. I explained that that was the acoustic tachometer ... you know that look you get ...
This bit discusses working memory and decision making (Page 25):

Actively holding information online for a brief period of time (seconds) is an important ability of the brain. This capability is a part of working memory (WM), which is used for tasks such as planning, organizing, movement preparation, and decision making.281–284 In contrast to long-term memory, which requires structural changes in neural circuits and in the connections between neurons, the mechanisms underlying working memory are believed to depend on persistent neuronal activity.285–287 In general, positive reverberation driven by recurrent synaptic excitation in interconnected neural clusters can work as the basic principle for generating persistent activity. Triggered by incoming signals, working memory circuits can sustain an elevated firing even after the inputs are withdrawn. As discussed in Sec. VB1 [see, e.g., Fig. 6(a)], dynamical models with local feedback excitation between principal neurons that are controlled by global feedback inhibition can exhibit multiple attractor states (each coding a particular memory item) that coexist with a background (resting) state. This is illustrated in Fig. 11 for a circuit model of WM with two excitatory populations.

In Akida 2000, LSTM will be a static equivalent of the dynamic "positive reverberation driven by recurrent synaptic excitation. ".

"In addition to robust maintenance of memory states, activity should be reset quickly when there is a novel sensory cue that needs to be stored. In other words, a working memory system should have the properties of robustness against fluctuations while being very sensitive to incoming stimuli. In recent theoretical works, this fundamental contradiction can be achieved by global inhibitory connections, where the system can exhibit structurally stable dynamics with fixed stimulus and qualitatively change its dynamics if the stimulus is changed."
I learnt a long time ago that academics proposing to provide a review of an entire area of science are having a lend and just trying to keep up their academic publication numbers which are apparently important in CV's for promotion in their world.

Poor old Peter van der Made would not get a job based on publications at the moment as he has been in stealth mode for years now.

I only posted the extract as being the easy read as I have searched for ages for someone prepared to encapsulate the outcome of understanding and creating a simulation of a cortical column. As for the maths I always assume that is correct and peer reviewed as I could not check it anyway.

The philosophy of whether we will ever understand the human brain using the human brain to derive an understanding and allowing for the variation in human brains and ongoing evolution is an intellectual quandary that I often mull over and cannot resolve. It does seem the more those engaged in this endeavour learn the less they know.

Personally I think that the human brain is a flawed model upon which to build useful artificial intelligence. The human brain makes mistakes constantly hence the saying "To err is human to forgive divine" or something like that. Being human not sure if I am misquoting but I will quote it nonetheless as I am using a flawed human brain to construct this response.

What use is an intelligent machine that perfectly mimics the human brain down to the making of inexplicable mistakes and the ego to keep going at it regardless.

In Autonomous EV's the push is in large part so that they can remove human drivers and eliminate accidents. Go figure that one out when you read about attempts to emulate the human brain to create truly autonomous vehicles. I am not Einstein but the logic breaks down in my opinion.

In short the artificial general intelligence that Peter van der Made is seeking must at some point divert from the position of fully replicating the human brain and as such a complete understanding of the human brain and how it works is unnecessary to this goal in my opinion but what would I know I am human after all and constantly right and wrong unless you ask my wife who will tell you I am constantly wrong unless I agree with her unless it comes to the law and investment and doing our tax planning.

My opinion only DYOR
FF

AKIDA BALLISTA


On-chip learning goes a way to meeting the requirement to store novel sensory cues.
...
Conclusion (Page 30)
One of the most relevant implications of the fact that cortical activity evolves as a sequence of discrete, metastable states is that transitions in neural activity are not just triggered by external events, such as a stimulus or a reward, but are instead spontaneously generated and may occur at anytime, including when the subject is idling and not engaged in a task. This goes against the notion that neural activity is just a “reaction” to external events or a static representation of incoming stimuli and is compatible with the presence of incessant “ongoing activity” observed in the cortex.

As FF asks, to what extent is it necessary to reproduce daydreaming?
 
  • Like
  • Love
Reactions: 13 users

mkg6R

Member
Hey guys tomorrow is a fresh month. Anything in particular thats on the cards for brainchip this month?
 
  • Like
  • Fire
Reactions: 3 users

Diogenese

Top 20
My immediate thought was the difference between the amount of concentration I give to your posts and the rocket memes of Rocket. LOL

I actually get what you are saying strangely enough.

One of the things I have wondered about is monitoring for noise and or vibration. Your sportscar analogy reminded me of it. Most of us have had the experience of someone new such as your young lady getting into our car and a couple of hundred metres down the road being asked what is that knocking sound or tap or something. Replying I can't hear anything out of the ordinary only to find later that some part had been wearing away over a long time and it was the cause of the noise detected by the young lady and not noticed by ourselves.

I have explained this away in the past to myself on the basis that it occurred so gradually over time that I became used to this noise. So the question I have is why would this not happen to an intelligent device monitoring for noise or vibration in a motor? The answer is I assume that there must be some way to set it to remember a perfect state and that any variation or noise outside that perfect state will be what it detects.

Now herein is my question. As no two mechanical devices will be in perfect harmonic balance one with the other it seems to me you would need to have an intelligent device that can learn that perfect state on the factory floor so to speak for the particular piece of machinery as it would be impossible to program every piece of individual piece of machineries noise or vibration perfect state in a central memory. In other words you ideally need on chip training preferably one or several shot in nature. Do you disagree?

My opinion only DYOR
FF

AKIDA BALLISTA
Hi FF,

My understanding is that there would be a range of acceptable vibrations either side of the "ideal" vibration, and, of course, the acceptable vibrations would be dependent on the rotational speed of the device.

So providing the individual machine's earprint is within the acceptable range, it should pass muster. But, I hear you say, if one machine is in the upper part of the acceptable range, and another machine is in the lower part of the acceptable range, is there not a greater chance of missing when one or the other approaches its damage threshold. And of course you would be right to be so concerned. So that may well be where the on-chip learning comes in to adjust the acceptable bandwidth to the individual machine by determining the "centre" about which the allowable variations could occur.

Or maybe, its just that the change from acceptable to "damaged" is so great that individual variations don't matter. After all, in the old days, an experienced mechanic could press his ear to a strategically placed screwdriver in the manner of a stethoscope to diagnose unhealthy rattles.

And, since we are moving to electric vehicles, there will be fewer bearings to monitor.
 
  • Like
  • Love
Reactions: 12 users


Technology's Legal Edge

A Global Technology Sector Blog

2024: The year that the future arrives​

Firework_Ligths_P_0141-X2

By Gareth Stokes on March 17, 2022
Posted in Artificial Intelligence, Hardware

Excerpt.

Sticking with silicon, various companies are exploring ways of delivering lower power solutions that will allow AI in a much wider range of devices. Two notable examples are Brainchip’s akida processors and Mythic’s analog (or ‘analogue’ in British English) AI processors. Brainchip’s akida takes a ‘neuromorphic’ approach, mimicking the spiking of neurons in the human brain. By contrast, Mythic’s chips take advantage of the electronically simpler designs for multiplication and addition operations in the analogue domain, for a different approach to low power AI inference solutions.
 
  • Like
  • Fire
Reactions: 24 users

JoMo68

Regular
This bit discusses working memory and decision making (Page 25):

Actively holding information online for a brief period of time (seconds) is an important ability of the brain. This capability is a part of working memory (WM), which is used for tasks such as planning, organizing, movement preparation, and decision making.281–284 In contrast to long-term memory, which requires structural changes in neural circuits and in the connections between neurons, the mechanisms underlying working memory are believed to depend on persistent neuronal activity.285–287 In general, positive reverberation driven by recurrent synaptic excitation in interconnected neural clusters can work as the basic principle for generating persistent activity. Triggered by incoming signals, working memory circuits can sustain an elevated firing even after the inputs are withdrawn. As discussed in Sec. VB1 [see, e.g., Fig. 6(a)], dynamical models with local feedback excitation between principal neurons that are controlled by global feedback inhibition can exhibit multiple attractor states (each coding a particular memory item) that coexist with a background (resting) state. This is illustrated in Fig. 11 for a circuit model of WM with two excitatory populations.

In Akida 2000, LSTM will be a static equivalent of the dynamic "positive reverberation driven by recurrent synaptic excitation. ".

"In addition to robust maintenance of memory states, activity should be reset quickly when there is a novel sensory cue that needs to be stored. In other words, a working memory system should have the properties of robustness against fluctuations while being very sensitive to incoming stimuli. In recent theoretical works, this fundamental contradiction can be achieved by global inhibitory connections, where the system can exhibit structurally stable dynamics with fixed stimulus and qualitatively change its dynamics if the stimulus is changed."



On-chip learning goes a way to meeting the requirement to store novel sensory cues.
...
Conclusion (Page 30)
One of the most relevant implications of the fact that cortical activity evolves as a sequence of discrete, metastable states is that transitions in neural activity are not just triggered by external events, such as a stimulus or a reward, but are instead spontaneously generated and may occur at anytime, including when the subject is idling and not engaged in a task. This goes against the notion that neural activity is just a “reaction” to external events or a static representation of incoming stimuli and is compatible with the presence of incessant “ongoing activity” observed in the cortex.

As FF asks, to what extent is it necessary to reproduce daydreaming?
The musings of my day job... working memory, planning and organisation, decision-making...love it! Working memory was once considered a proxy measure for intellect...
 
  • Like
  • Fire
Reactions: 7 users

Esq.111

Fascinatingly Intuitive.
Hey guys tomorrow is a fresh month. Anything in particular thats on the cards for brainchip this month?
Afternoon mkg6R,

For the month of April I only have three events penciled in so far,

1, April 2 , Pod cast, Chit chat with Chairman of Foundries.io.

2, April 17, OTC Markets, Technology Virtual Investor Confrence, BRN presenting.

3, April 29, 4C, Quarterly Report, ( possibly disclosing multiple billion $ contracts ) , * That last bit I added, unlikely but possible.

& today , March 31, I thought we were going to receive the Audited Accounts.???

BRN Shareholders need some, Positive reverbiration driven by recurrent synaptic exitation.

I have stolen this bit from one of Diogenese posts.

Regards,
Esq.
 
  • Like
  • Fire
  • Love
Reactions: 39 users

McHale

Regular
Hi FF,

Now you've triggered one of my pet peeves.

This is the biography of "those who without any connection to the real world (where we all live) are able to manipulate politicians, regulators, the media":
Davos Man
How the billionaires devoured the world
by Peter S Goodman
https://www.dymocks.com.au/book/davos-man-by-peter-s-goodman-9780063239340
They're not very nice people, but IMO Klaus Schwab (WEF) is the main man, piece of work.
 
  • Like
  • Fire
Reactions: 9 users
My immediate thought was the difference between the amount of concentration I give to your posts and the rocket memes of Rocket. LOL

I actually get what you are saying strangely enough.

One of the things I have wondered about is monitoring for noise and or vibration. Your sportscar analogy reminded me of it. Most of us have had the experience of someone new such as your young lady getting into our car and a couple of hundred metres down the road being asked what is that knocking sound or tap or something. Replying I can't hear anything out of the ordinary only to find later that some part had been wearing away over a long time and it was the cause of the noise detected by the young lady and not noticed by ourselves.

I have explained this away in the past to myself on the basis that it occurred so gradually over time that I became used to this noise. So the question I have is why would this not happen to an intelligent device monitoring for noise or vibration in a motor? The answer is I assume that there must be some way to set it to remember a perfect state and that any variation or noise outside that perfect state will be what it detects.

Now herein is my question. As no two mechanical devices will be in perfect harmonic balance one with the other it seems to me you would need to have an intelligent device that can learn that perfect state on the factory floor so to speak for the particular piece of machinery as it would be impossible to program every piece of individual piece of machineries noise or vibration perfect state in a central memory. In other words you ideally need on chip training preferably one or several shot in nature. Do you disagree?

My opinion only DYOR
FF

AKIDA BALLISTA
Hi FF,

My understanding is that there would be a range of acceptable vibrations either side of the "ideal" vibration, and, of course, the acceptable vibrations would be dependent on the rotational speed of the device.

So providing the individual machine's earprint is within the acceptable range, it should pass muster. But, I hear you say, if one machine is in the upper part of the acceptable range, and another machine is in the lower part of the acceptable range, is there not a greater chance of missing when one or the other approaches its damage threshold. And of course you would be right to be so concerned. So that may well be where the on-chip learning comes in to adjust the acceptable bandwidth to the individual machine by determining the "centre" about which the allowable variations could occur.

Or maybe, its just that the change from acceptable to "damaged" is so great that individual variations don't matter. After all, in the old days, an experienced mechanic could press his ear to a strategically placed screwdriver in the manner of a stethoscope to diagnose unhealthy rattles.

And, since we are moving to electric vehicles, there will be fewer bearings to monitor.
Thanks once again.

I had a mate who would tune the twin su’s on my sports car with a piece of hose stuck in his ear and the other end held in front of each air intake and he would move the hose back and forth until he had them sounding exactly the same.

I wonder if Robot Ken could do this. 😂
FF
 
  • Like
  • Haha
Reactions: 11 users
Afternoon mkg6R,

For the month of April I only have three events penciled in so far,

1, April 2 , Pod cast, Chit chat with Chairman of Foundries.io.

2, April 17, OTC Markets, Technology Virtual Investor Confrence, BRN presenting.

3, April 29, 4C, Quarterly Report, ( possibly disclosing multiple billion $ contracts ) , * That last bit I added, unlikely but possible.

& today , March 31, I thought we were going to receive the Audited Accounts.???

BRN Shareholders need some, Positive reverbiration driven by recurrent synaptic exitation.

I have stolen this bit from one of Diogenese posts.

Regards,
Esq.
Or was it the BeeGee’s? 🤣 FF
 
  • Haha
  • Like
Reactions: 4 users

mrgds

Regular
Now that is one BOLD statement FF,... and to be more specific, .... " Brainchip is on a trajectory to success that if achieved will be unlike anything ever seen before in the commercial and scientific world."

That's triple expresso bold. Not that I disagree necessarily, but I am less confident on that outcome, let us say. I'd be happy with just the Nasdaq 100.

Fortunately, you left an escape exit by use of a qualifying " if...". Sneaky,.... You sure do craft wonderful positions usually supported with good proof or evidence. Keep up the great work and be as colorful as you wish. Sometimes you crack me up, though,....

Or was it the BeeGee’s? 🤣 FF
Maybe Beachboys ? ................ Good "vibrations", and "exitations"
 
  • Haha
  • Like
Reactions: 13 users
They're not very nice people, but IMO Klaus Schwab (WEF) is the main man, piece of work.
An interesting insight into your main man. LOL:


FF
 
  • Like
Reactions: 3 users

equanimous

Norse clairvoyant shapeshifter goddess
Thanks once again.

I had a mate who would tune the twin su’s on my sports car with a piece of hose stuck in his ear and the other end held in front of each air intake and he would move the hose back and forth until he had them sounding exactly the same.

I wonder if Robot Ken could do this. 😂
FF
FF we just need your tuning requirements please. We are working on it
 

Attachments

  • Robot and car.png
    Robot and car.png
    3 MB · Views: 57
  • Like
  • Haha
  • Love
Reactions: 7 users
Top Bottom