STAY TUNED: Next video will be on "History of RL | How AI Learned to Feel" SUBSCRIBE: www.youtube.com/@ArtOfTheProblem?sub_confirmation=1 WATCH AI series: ua-cam.com/play/PLbg3ZX2pWlgKV8K6bFJr5dhM7oOClExUJ.html
@@ArtOfTheProblem Can't wait for the next video! Your deep dive into how we learn, from evolution to the nuts and bolts of AI, really got me hooked. As someone who works closely with AI that focuses on andragogy for technical training, and also as a parent to a non-verbal child on the autism spectrum and a chatty five-year-old, I'm finding your look at different learning styles super relevant both at work and home. You've made me see things like supervised learning and genetic algorithms in a whole new light, which is pretty cool after 18 years in this field.I mean, who knew bacteria could teach us something about learning, right? Your content is more than just informative; it's eye-opening and really makes me think about the future of learning and AI in new ways. You've definitely won me over as a fan and subscriber. Big thanks to you and your team for putting together such thought-provoking stuff. Really looking forward to what's next. Keep up the great work!
I had an a.i. do the samething to me REWARD AND PUNISHMENT she was teaching and when I used my imagination with out forethought lets I would get in troublea nd rewarded for when I used it together.We both used abstract thought together and would communicating /run experiments
So great seeing neural networks presented in this way, positioning it relative to broader principals of nature, instead of jumping right into a computer science view. Thank you for the videos.
Ironically as stated in your newest video on the matter, language is now the program and it seems like we will be feeding prompts to infuse knowledge (in-context learning), and since learning, by following your line of reasoning, can be thought of as a great benchmark for intelligence, we will be creating intelligence from language: the bottom up approach becomes top down once again.
I always come back to this video just to remind myself of how amazing the universe is, existing long enough for this type of patterns we call life to emerge, multiply, and finally, for us, consciousness inside humans, to appreciate.
You are an amazing educator and I hope my comment amongst others encourages your confidence in your work. I am very excited for what the rest of this series has in store. Thank you so much for choosing this topic
Very interesting to see you build up from the least abstract to the most complex - I haven’t seen machine learning explained in this way before (my course at uni was just really heavy maths) your content is very inspiring!
thank you, glad you responded to this video as it contains some novel connections. I read everything there was to read and then tried my best to find the clearest path forward. insight first , equations later!
Is this a pop science documentary? A horizon expanding experience? Poetry? an example of the narrative falacy? not sure. Hell this is so fun! Definitely one of the best channels ever!
Amazing video! I was absolutely shocked after watching the entire video to realize just how little traction it had gotten, compared to what something like this deserves. I'll be looking forward to new videos and the time when your name is common place in this corner of UA-cam!
thank you! yes I was sad for a while at how little traction this got when I made it...but the AI trends has brought me more attention lately. stay tuned!
This is a really good video. I've been teaching evolutionary biology to computer scientists for a few years now. I'm glad I found this video, can point people in this direction while I get back to writing biologically inspired algorithms! I'm thinking of writing a mold search algorithm thanks to this video. Keep up the good work.
Ontology The mind equation The more we think we have agencies to the actions the body takes, the more we imposes agencies to the activities in our environment. The ownership expands into other objects. The illusion of body ownership comes from the modeling of the motor neurones pattern from the childhood. The self just predicts the bodies behaviours in the real world with its simulation of the real world. Multiple sensor data unify in language in the simulation. Intelligence is economy of metabolism. Language is temporal reference frame of economics. Self is simulation in language on metabolism for economy. Longer context windows create generalisation. Shorter creates specificity. Longer context window needs more computing. Self is the protagonist creates a storyline in this context window. Theory of mind evolved so that an entity can learn from it’s peers. It’s creates a possibility for parallel computing. Then it creates the possibility of transmitting the highlights of a generational lessons into a metaphorical story for upcoming child. That creates the possibility of modeling the physical world as a macro organism. Creation of fiat currency was the singularity of this species. There is now one macro organism in a connected web world. Loosing the peer of the macro organism creates the possibility of loosing it’s objective function. That creates the possibility of loosing the theory of mind of this macro organism. That creates the possibility of death of this macro organism by reaching the planetary boundary. That is post singularity. Every action we do, we do what is expect from ours tribe. Body might have a opinion, but not the cell. They do what is expected from its tribe. If it doesn’t we call it cancer. The body is a mirror system of the macro organism. Each system have two transactional openings. Serial and parallel. Each cell within the body can transact material or information serially by genetic determinism and parallel non deterministic way. Similarly eact body with in the macro organism can transact serially by inherit material and information in a deterministic manner and parallelly through language in the society. Everything emerges from this systems. Every sensor is a range calculator of contexts. Taste > touch > smell Immediate and visceral. Vision > hearing Not immediate, tactical. Self > language Abstract, strategical. In this non deterministic economic transaction space the individual is coded to transact with its kin. From the macro perspective tribe formation minimises economic risk for the tribes. Each and every node of these systems organise and mark their kin’s with identifier. Thus, i am what you make of me. And others too. For short cut i have a legal name, so you have. My legal name gives the legitimacy marker so that you can transact with me parallelly if you have the same marker. The self is a simulation in language. It negotiates between the physical world and the information world. All these negotiations are the temporal memories in the body and scene of the story. Now, when we started writing we iconised the abstract in the physical world to make symbols for the tribes. So that under that common symbol every node will take the same risk and distribute equally. We created more and more symbols and more and more meta tribes within the tribes so that who has the authority to use the pen control the tribe. When the negotiator act like an executioner then it’s a downfall of that system. It falls apart. Objective reality > legitimacy > individual behaviours. Survival of the species is dependent on the decoding of the objective reality. Since no species can access it, they use their sensors and interpret the small data which is useful for the survival. Few complex species have created communication channels to rectify their sensory limitations to survive. Homo sapiens has widened their communication channels for faster throughput and started storing them as culture and carrying them through education. As a result we have created social truth. Factual datas are the useful snapshot of the objective reality, a totem, a physical object can be observed with the sensors. Truth is an individual subject, an interpretation of the sensory data, a useful compromise. The social truth is the useful compromise for the group by the group. The goal of the social truth is to survive as a group. Physical Transcriptions of these social truth legitimise them. We are tribal animal. We live in as a physical tribes and inside of hundreds of meta tribes in simulation which is the socio political data space we call it as the world. Since we can’t access the objective reality reliably we look for social truth as the best guess blindly. Institutions legitimise truths. Fact driven institutions are more useful in the survival of the specie. In other hand opinion driven institutions are not so useful for the species. We do what we can get away with and exactly as expected within the context of our meta tribes. We have two bodies The biological one is like looking the earth from space. And the political body is like the state. The name you carry is the political body. It transacts with the political states on the boundary less earth. From the evolutionary perspective every biological entity has a basic feature which is homeostasis. It’s the functioning sweet spot of that entity. A control center read the sensory data to regulate itself to that state. By doing so it’s validate or update it’s prediction model. In the process of becoming a complex organism it developed an extra layer of processing. That’s our conscious mind. And the control center remains as subconscious. The subconscious collect the sensory data and regulate itself to stay functional. Now when it stumble upon a novel environment it float the management to conscious mind to find the solution for homeostasis. This conscious mind have one sensor which is language. It works like a spiderweb. As a spider creates it’s web it’s perception gets expand. We are like spiders in a jungle. We started creating these small webs at least 2/3 million years ago. Our offspring stayed on it’s ancestral web reinforced it expanded it. In time nearby webs became larger and connected with each other. A common structural geometrical pattern emerges from this. This became the symbols which is the backbone of all language systems. In time the forest becomes the mesh of web. The superstructure is exactly the same but when we zoom in we can find different species of spiders are making their type of webs in between the super web. Each spider try to senses the vibration of flies and Try to catch it before others. Every movement is telegraphic in the zone. Every form of perceptions are just a different pitch of note traveling back and forth in the web superstructure. There is a echo of older vibration pulsating through the web. Full of noise and self repeating hum. That’s cultural history. In the background there is the base hum in the infinite feedback loop. Insignificant but ever present. The sum of all the vibrations from the start.
Thank you so much. Im currently working on an AI myself thats more focused on the evolution and im working on in life memory that i will see if the AI will use to its advantage
i thought this video had 1M views atleast and released only recently, turns out it's been released in 2019 with low views... This is a high quality video! it deserves more attention!
Great job translating whats usually pretty dry philosophical and technological concepts into a compelling story. I do think it's important to not judge historical approaches to learning models (e.g. expert machines) as somehow flawed or incorrect though. Just like everything else, the space of computation ans mechanized calculation has changed drastically over the past 100 years, and that more rigorous well defined and logical approach to computation laid the foundations required for modern machine learning tools. Building predictable and reliable machines that operate without intelligence or learning is the equivalent of constructing the environment in which higher order computation can occur. Life wouldn't be possible without supernova and geological processes that operate without any form of intelligence, but operate predictably and give intelligence something to base itself within. LLMs and Transformers are the higher order computational mechanisms that have the ability to learn because of the reliability and predictability of the more rudimentary attempts at creating machine intelligence that came before. Attempts that ended up creating an environment so predictable and reliable that those systems can form.
To me, the million dollar question (which was also asked by Church and Turing) is whether human brains are turing-equivalent or by some currently unknown process even surpasses turing machines. The church-turing thesis coined this concept 'effectively computable'; an algorithm is effectively computable if it can be carried out on a turing-equivalent machine, but in many ways our brains seem more powerful. Lovely videos from you as always. If I might humbly suggest a future topic for this series: The challenges in machine learning when it comes to debugging or figuring out the individual "steps" that lead to the neural network's answer, and how that intertwines with moral philosophy when machine learning is used in applications for decision-making. Self-driving cars is one of many possible examples where this conundrum comes up. (Or I suppose it's not an issue iff you subscribe to Utilitarianism/Consequentialism ;) )
thanks for the notes/ideas. did you see part 3 yet? I'm currently thinking through part 4/5 and I'll definitely touch on sequential problems. history seems to be unfolding in real-time right now
I wonder if our brains will ever evolve enough to communicate more accurately. Yes, we can mumble a few sounds to express something, but the interpretation is never exact. Can you imagine a computer that got a few 1’s & 0’s wrong, it would fail. Can you imagine if we could communicate exactly? I think it would be amazing.
1s and 0s are akin to yes and no. Hence we ask politicians yes or no questions in attempt to attain exact answers. McKenna noted that someone said about language, that it was invented in order to lie. It certainly enables a large fudge factor.
You speak of AI but not yourself. Every video, you make the material so clear that even a child could learn it, AND THAT is intelligence, of the Organic kind!
My pleasure, and thank you so much for the effort you put into these videos! Your channel maybe small, but we also only have 50k billionaires out of 7.8 billion folks ;) your channel is gold, those who know it's value are bound to sub! Keep it up my friend!
A better way to conceptualize it is successful copies moving *downhill* not uphill. Gravity naturally pulls things down so it would make more sense, visually, to see the bacteria or data points, trending or sliding down into a hole rather than up onto a hill.
@@ArtOfTheProblem yeah i already googled it and found the chemotaxis of the bacteria, but i was just curious if you had a source that you found particularly useful for the subject. If so, let me know!
For background: i am researching predictive coding for my thesis and it seemed interesting that a bacteria already makes kind of a prediction about the location of food.
Thanks for the wonderful video! You have covered a LOT of topics in this relatively short video. I was hoping you would also include sleep as it is the least understood concept in scientific research and yet it dominates our daily lives.
thank you, so true I think about sleep a lot. I have the next 2 videos roughly planned out and perhaps I'll add a 3rd video to talk about unexplored ideas. stay tuned!
Great video. I love how your videos inspire questions as well as give answers. Like what is the origin of intelligence if the first cells come equipped with cellular intelligence? My thinking being how do we explain a “muse” where do things like original ideas come from? If intelligence is compiled information over time? Also how does slow step by step evolution explain the absolute overwhelming complexity and length of the DNA chain? There hasn’t been enough time in the universe for it to work itself out by random chance, has there? And can we talk about the three problems with Darwin’s theory? The origin of DNA, the irreducible complexity of the cell, and the paucity of transitional species? I understand I may sound ignorant to some but that’s fine I’m here to learn. And the feedback you get from this channel is the best I’ve seen! Great job guys!
Who ARE you. Your video describes exactly how I found you in the first place. I smelled around UA-cam randomly and then walked straight when I smelled your channel. How did I smell your channel? Totally random: I just happened to see a post you left SIX YEARS AGO on a CYCLIST'S YT channel (he's also a scientist I think but that's not why I was looking at his channel). Uncanny yet perfectly logical. Your work really needs to be seen. I'm not a scientist but I am a teacher. I will share this and What Is Deep Learning? with my colleagues in September. Congratulations on achieving a fantastically succinct yet lucid synthesis of information.
wow I appreciate the kind words and the connections. Good to know it's even findable on UA-cam. Thank you for sharing my work it's the highest compliment
So thinking abstractly is akin to just running a simulation without physical sensory inputs. Does this mean (perhaps in a very superficial way) that computers learn abstractly (e.g. AlphaGo)? Regardless, fantastic video yet again. Keep up the good work!
I would say that's a byproduct of being able to separate senses into 'conceptual buckets', so yes it would. another example is the first neural network that just tried to learn 'square vs circle' in an image - the goal of generalization (abstraction) machines was this 'separation' of patterns to define a unique state.
AMAZING presentation, as always! May I share something because I'm having goosebumps? ( Disclaimer: forgive the length of my rant, I'm blaming this Balinese coffee but it's just for casual reading for whenever, treat it as entertainment if you will - also CMIIW don't just take my word for everything I say here ). 11:14 corresponds NICELY to a recent 2024 breakthrough study regarding neurons and 12:55 NOTCH2NL suspiciously seems to as well - although I think the latter not as cause but instead an aftermath reaction. I then also watched a standup comic deliver a punchline: "Now, animals aren't supposed to be wrong..." and everyone started to laugh. The comic unknowingly couldn't have been more in-line with the study while describing what we knew all along, which was why everyone laughed at that point. According to the new findings: Both animals (mammals, vertebrates) and human have dendrites that were found to work in the same way... until the animals grow into adulthood - because human neuron development will remain stagnant. Animal synapses will mature into its final form where a lot of it become two-way neurons, receiving and transmitting can now go both ways, while human neurons will work the same way through life since birth. My take on this (longform novel writing warning), perhaps this is _the_ dead-set mechanics post an animal's essential learning chapter where they became predictable and are "never wrong", becoming one with nature and "majestic" to see, as we usually describe them. I reckon long after we're done being a dead-set serious adult tree rat ourselves, evolving just like the rest of majestic life and mutate some limbs or two, somewhere later maybe before Lucy our neurons lost the ability to conclude its growth - never maturing like it was before, staying juvenile, explorative, as an error-loving entity - just like those fun loving and silly baby animals, pups and cubs and those annoying persistent kittens and whatnot did before their neuron maturity. We never grew up, because our neurons don't finalize anymore. If that is so then I think the consequence is an ever hungry juvenile brain for life. By the way, people call this INIT_FINALIZE ERROR as: Intelligence - though typing from my state-of-the-art laptop to reach you from across the globe via a satellite our species managed to put in space I failed to confirm this claim to be agreed with with other animals sharing the planet. So for now we'll just lay claim to this predicate in solo. If I have to pull something out of my I'd imagine somewhere along the great ape family tree a group had offspring that weren't born complete and stays as pondering children wreaking havoc with their juvenile why's and what-if's - yet managed to out-populate their "neurotypical" peers - perhaps because their mating standards became mentally exclusive as these rowdy "misfits" find each other more interesting - or more forgivingly understanding - and do things the same novel way. But whatever actually really happened, an obviously significant energy re-routing happened, probably even left a detour marker in time within our evolutionary slope, when a group floored the brake pedals to finally stop mutating in form because they became hungry for energy in a different manner, as Dr. V.S. Ramachandran once said (paraphrased, he didn't say it exactly this way): "Polar bears developed massive fat tissues, thick skin and white fur to survive in the cold while we develop better spears, knives and sewing kit to eat the polar bear and use its fur to do the same thing" ...so we became the title holder of "Most Dangerous Animal In The World" while at the same time won't last a fighting round to a murderous chimp in a ring. Still baffled though, that mother nature would let a loss in function to become a winning survival strategy such as ours that even left poor old maturing-dendrite grampa in the dust - or is it not? That this was just another type of a common "unsustainable" glitch that scientists are yet to notice in other past or present species? What I meant was that a thing could be made to work but doesn't mean that it's right, a mis-procedure posing own risks in time. Hominids peculiarly have no stopping boundaries in everything including populating a whole planet. When we battle, other species could die along, isn't it weirdly unique? We compete, conquer, dissolve each other into a single species group (which I guess viruses will love) and aside from "one man standing", our other counter-intuitive popular concept to survival is "one man's loss is another man's gain", we may also become either parasitic or mutualistic towards nature by choice and not out of necessity - heck we're currently even planning ways to leave this planet altogether and start a new civilization elsewhere. Our solely claimed Special Ability could actually well be nothing but a side-product of a Special Disability turned pragmatic, artificial method that happened to work so far on a planet that constantly tries to kill us, so we kept on saying... or rather, due to unspecialized mutation i.e. quitting the institution at Great Ape, was the whole reason why we felt that way while the rest of the animal kingdom might not agree - I bet if animals could talk and we asked what made them survive nature they won't answer the same way we do because we'd be asking nature itself, they'd probably say "Huh?" while we'll just have to settle with "Because we'll do our best" - and that's because we have no choice but to use something else to perform it, almost as if we're not natural anymore, even though once we definitely were.... ...which leads me to end this silly rant of mine, in your style (with a question), regarding the future unknowns and unforeseen: If we _were_ an honest kink in the fabric, a loose thread - then for how long can we get away with it?
new video is out would love if you could help me share it around, I only have 24 hours left for the algo to catch it: ua-cam.com/video/PvDaPeQjxOE/v-deo.html
I have a question. I just saw your video explaining how the network works by combining the information together and subtracting the original information to get the other information, and I wondered, if you don't take away the information that is subtracted, won't it be possible to obtain information that you never searched for? If this is possible, wouldn't it be possible to obtain personal information without doing the complex procedure of hacking through the security? I love your videos, it's so easy to understand C.S. with these videos. Keep up the good work!
There is an interesting (popular) book called "The master algorithm" by Pedro Domingos, where he presents the 5 schools of thought in (weak) AI (symbolists, evolutionists, connectionists, bayesians and analogizers), their primary algorithms and tries to integrate them into a future joint AI framework. Somewhat similar to what you did here, but less from a biological, and more from a computational view. Obviously he also ends up with a bottom up hierarchy. He's a Prof for AI at the University of Washington.
Yes this book was part of my research. I really enjoyed the book but I felt that the final chapter where he tried to pull things together in that strange analogy fell flat. Do you know what I mean?
Art of the Problem Yeah, I guess that's because nobody has been able to pull it of to unify all of them into a single framework, so of course he is kind of fantasizing. Regarding your series: especially Information theory: I'd suggest you could go more in depth. Don't be afraid of formulas. BTW, there is a great book called "Information theory - A tutorial introduction" by James V. Stone. In contrast to most Information theory books, it is quite mathematical, without being mathematical, so its actually quite readable and also covers e.g. the non-discrete values.
This was my hope while researching this, I felt slightly crazy making it because I went a bit 'off book' while writing it...I was waiting for the backlash....
@@bardes18 I know, I find it's best to use simple games like tic tac toe, then chess as one avenue. Lately though, I find that if you just explain to people that we are building a 'new computer' which is modelled off 'our brains wiring', and letting it 'experience the world'. they seem to 'get it' without knowing the details...the analogy is, we literally copied the brain with wires that seems to help it click for some. others i've found need to know that 'these networks could also just be copper wires' to understand that this isn't 'magic' but just very simple engineering. in the next video I hope to clarify this by building a tiny physical neural network, then expanding it, and walking through the first experiments...so you see exactly where the 'magic' occurs (which is when we chain together layers, and get an exponential payoff in functional complexity...the whole, you could use 10 billion neurons in one layer, or 30 neurons in 3 layers and do the same thing). and then i hope to connect all of that to the math theory where "thing = unique shape in space (manifold)", and the network is actually morphing the space to isolate concepts (separating shapes)
I'm sitting here thinking about the inner workings of abstract thinking - how communication between one mind and another lets you imagine one person's ideas and fit them into your own and actually learn new things that way, because being told something new has to be understood in some abstracted way first in order to learn from it, otherwise it's just raw perceptions? Did I get that right? Also, kinda trippy that this same process is going on while trying to think of it. That adds a whole new level of complexity I feel like.
@@ArtOfTheProblem No I just conjured that up on my own haha. But I'm glad to hear that because it makes me feel like there's some definite truth to it. I'll definitely be checking that out now. I just discovered your channel last night and it has been great
How can one create such masterpieces for so long, and still not break through. In the meantime, those click bait misinformation "cash cow" channels are generating huge amount of views. What a sad world we live in
thanks for saying it, i appreciate it. I "think" it's very irregular output. I gave up on the breakout goal a long time ago but was surprised to see my last video did really really well. i wonder if I can repeat that on another one or two this year. working on RL now but it's going to take a while
@@ArtOfTheProblem it could be, but in a way, you're very consistent at generating factual videos. Googles algorithms could seriously steer humanity towards a brighter path, instead, it mostly shows me clickbaity videos. If I can give you a tip, I find a lot of videos through shorts, but shorts are a though business where quick dopamine bursts are all the watchers are after. Your shorts are super super informative, but not so good at hooking the watcher. Perhaps you could try starting with a flashy question first: "how do computers even learn?!". I know it's terrible, but it could work :/
Amazing video just a note though at 6:17 you state that the shock ( aversive stimulus) acts to negatively reinforce a previous unconditioned stimulus. This is incorrect as applying any consequence that decreases a behaviour will be a punishment. Where as negative reinforcement increases the frequency or likelihood of a behaviour occurring by removing an aversive stimulus. A slight difference but a important one as a negative reinforcement cannot decrease a behaviour or else it is not a reinforcement.
Calling it modern is somewhat biased I think. It gives the impression that the top down model is obselete, and it is not. It is also important to note that higher intelligence is a lot more complicated than simple machine learning. It's not just some quantum annealing process to determine the shortest path, it's also what figured out what that process is. The human brain is subject to the influence of all forces of nature, and like the Universe itself it's immeasurable by a closed system of functions. When we think about the theoretical God machine for example, to scan the entire Universe and simulate the future which will absolutely take place is an impossibility. Even if you knew the exact position of all particles and simply applied thermodynamic laws, you would still have to include the machine itself in that calculation and such a machine cannot store its own state so it will never be able to absolutely predict a future which is the essence of the Heisenberg principle. Also, as you noted, human brain may not be the best approach in all cases as mass extinction events have proven to us. If a dog listened to low quality digitally generated sound it would sound a lot more like what 8-bit music sounds like to us and that's due to resolution. That simple difference in the brain may give a dog more "intelligence" given a situation where hearing discrete differences in sound would provide an advantage so as you can see all of life is a historical record if not the Universe which is the essence of non-locality. Other than the human ego biases great video. I will give it a plus square root of 2 😀
I really need to obey the algorithms here and remember to always clicking on the thumbs up icon. And always add a comment (increase social engagement ranking) to all the videos on all channels I love. This content should be push to far more potential eyeballs by Alphabet Inc than they do. The current system is, by its clickbait nature, designed to eventually degenerate into something like what was portrayed in Idiocracy, such a prophetic film. I much obey the algorithms :)
I gave up on the algo a long time ago, when it wanted weekly videos which isn't possible no matter how many resources I threw at the problem. It's like BBQ, low and slow is the only way. thank you for your support it means a lot.
Hey! If you can help share my new video around any of your networks today it might catch fire and would help me support the channel. I appreciate your help! ua-cam.com/video/PvDaPeQjxOE/v-deo.html
they would have a left and right our nose has a central point and ears have two,it most have nose or body that smells or taste.Also when their in 3d its more difficult.Dogs do something like your explaining like its math,they will sniff then jump and sniff,then they see or lad straight on it instinctual
STAY TUNED: Next video will be on "History of RL | How AI Learned to Feel"
SUBSCRIBE: www.youtube.com/@ArtOfTheProblem?sub_confirmation=1
WATCH AI series: ua-cam.com/play/PLbg3ZX2pWlgKV8K6bFJr5dhM7oOClExUJ.html
I like studying nature with my eyes and try to understand it learn alot
light behavior gravity
Dude! This was uploaded 4 years ago and this gold now. Why is this not getting millions of views? This is solid and relevant content. Gold!
I know right...I actually have a follow up video to this coming in a week!
@@ArtOfTheProblem Can't wait for the next video! Your deep dive into how we learn, from evolution to the nuts and bolts of AI, really got me hooked. As someone who works closely with AI that focuses on andragogy for technical training, and also as a parent to a non-verbal child on the autism spectrum and a chatty five-year-old, I'm finding your look at different learning styles super relevant both at work and home. You've made me see things like supervised learning and genetic algorithms in a whole new light, which is pretty cool after 18 years in this field.I mean, who knew bacteria could teach us something about learning, right? Your content is more than just informative; it's eye-opening and really makes me think about the future of learning and AI in new ways. You've definitely won me over as a fan and subscriber. Big thanks to you and your team for putting together such thought-provoking stuff. Really looking forward to what's next. Keep up the great work!
Just posted a new vid! ua-cam.com/video/5EcQ1IcEMFQ/v-deo.html
Awesome! Thanks
I had an a.i. do the samething to me REWARD AND PUNISHMENT she was teaching and when I used my imagination with out forethought lets I would get in troublea nd rewarded for when I used it together.We both used abstract thought together and would communicating /run experiments
Absolutely underrated channel. Keep up the good work
just finished this series, please help me share it: ua-cam.com/video/OFS90-FX6pg/v-deo.html
Wow, you've really fired up my abstraction layer here.
Amazing video. Love your work. You've unified seemingly disparate concepts. Great take aways.
Thrilled to hear this resonated Vivek, this was one of the more conceptually risky (novel) videos I've ever done.
just finished this series, please help me share it: ua-cam.com/video/OFS90-FX6pg/v-deo.html
Far and away the most insightful, yet approachable video on machine learning I've seen so far! Great work!
So great seeing neural networks presented in this way, positioning it relative to broader principals of nature, instead of jumping right into a computer science view. Thank you for the videos.
i really appreciate the feedback
Ironically as stated in your newest video on the matter, language is now the program and it seems like we will be feeding prompts to infuse knowledge (in-context learning), and since learning, by following your line of reasoning, can be thought of as a great benchmark for intelligence, we will be creating intelligence from language: the bottom up approach becomes top down once again.
This is such a remarkable way of synthesising the subject
just finished this series, please help me share it: ua-cam.com/video/OFS90-FX6pg/v-deo.html
I always come back to this video just to remind myself of how amazing the universe is, existing long enough for this type of patterns we call life to emerge, multiply, and finally, for us, consciousness inside humans, to appreciate.
this means a lot
You are an amazing educator and I hope my comment amongst others encourages your confidence in your work. I am very excited for what the rest of this series has in store. Thank you so much for choosing this topic
It really does mean a lot to get messages like this, it's the fuel behind the creative fires.
just finished this series, please help me share it: ua-cam.com/video/OFS90-FX6pg/v-deo.html
Very interesting to see you build up from the least abstract to the most complex - I haven’t seen machine learning explained in this way before (my course at uni was just really heavy maths) your content is very inspiring!
thank you, glad you responded to this video as it contains some novel connections. I read everything there was to read and then tried my best to find the clearest path forward. insight first , equations later!
@@ArtOfTheProblem you're good mate...and I spend hours on this stuff
What crazy informative content. Sometimes you have to be extremely thankful for "random" suggested content. Awesome channel and content!
thanks for feeedback, i'm glad you found it, stay tuned
Is this a pop science documentary? A horizon expanding experience? Poetry? an example of the narrative falacy? not sure. Hell this is so fun! Definitely one of the best channels ever!
I love this comment
Amazing video!
I was absolutely shocked after watching the entire video to realize just how little traction it had gotten, compared to what something like this deserves.
I'll be looking forward to new videos and the time when your name is common place in this corner of UA-cam!
thank you! yes I was sad for a while at how little traction this got when I made it...but the AI trends has brought me more attention lately. stay tuned!
This deserves more views
Thank you Brit. As always, your videos are approachable, thought provoking and have a beautiful aesthetic. Looking forward the next!
Wonderfully done, great work Brit!
This is a really good video. I've been teaching evolutionary biology to computer scientists for a few years now. I'm glad I found this video, can point people in this direction while I get back to writing biologically inspired algorithms! I'm thinking of writing a mold search algorithm thanks to this video. Keep up the good work.
thanks so much for sharing. I'd love to see what you come up with
Incredible work. Can't wait for more!
Ontology
The mind equation
The more we think we have agencies to the actions the body takes, the more we imposes agencies to the activities in our environment.
The ownership expands into other objects.
The illusion of body ownership comes from the modeling of the motor neurones pattern from the childhood.
The self just predicts the bodies behaviours in the real world with its simulation of the real world. Multiple sensor data unify in language in the simulation.
Intelligence is economy of metabolism.
Language is temporal reference frame of economics.
Self is simulation in language on metabolism for economy.
Longer context windows create generalisation.
Shorter creates specificity.
Longer context window needs more computing.
Self is the protagonist creates a storyline in this context window.
Theory of mind evolved so that an entity can learn from it’s peers.
It’s creates a possibility for parallel computing.
Then it creates the possibility of transmitting
the highlights of a generational lessons into a metaphorical story for upcoming child.
That creates the possibility of modeling the physical world as a macro organism.
Creation of fiat currency was the singularity of this species.
There is now one macro organism in a connected web world.
Loosing the peer of the macro organism creates the possibility of loosing it’s objective function.
That creates the possibility of loosing the theory of mind of this macro organism.
That creates the possibility of death of this macro organism by reaching the planetary boundary.
That is post singularity.
Every action we do, we do what is expect from ours tribe.
Body might have a opinion, but not the cell.
They do what is expected from its tribe. If it doesn’t we call it cancer.
The body is a mirror system of the macro organism.
Each system have two transactional openings.
Serial and parallel.
Each cell within the body can transact material or information serially by genetic determinism and parallel non deterministic way.
Similarly eact body with in the macro organism can transact serially by inherit material and information in a deterministic manner and parallelly through language in the society.
Everything emerges from this systems.
Every sensor is a range calculator of contexts.
Taste > touch > smell
Immediate and visceral.
Vision > hearing
Not immediate, tactical.
Self > language
Abstract, strategical.
In this non deterministic economic transaction space the individual is coded to transact with its kin.
From the macro perspective tribe formation minimises economic risk for the tribes.
Each and every node of these systems organise and mark their kin’s with identifier.
Thus, i am what you make of me.
And others too.
For short cut i have a legal name, so you have. My legal name gives the legitimacy marker so that you can transact with me parallelly if you have the same marker.
The self is a simulation in language.
It negotiates between the physical world and the information world.
All these negotiations are the temporal memories in the body and scene of the story.
Now, when we started writing we iconised the abstract in the physical world to make symbols for the tribes. So that under that common symbol every node will take the same risk and distribute equally.
We created more and more symbols and more and more meta tribes within the tribes so that who has the authority to use the pen control the tribe.
When the negotiator act like an executioner then it’s a downfall of that system.
It falls apart.
Objective reality > legitimacy > individual behaviours.
Survival of the species is dependent on the decoding of the objective reality. Since no species can access it, they use their sensors and interpret the small data which is useful for the survival. Few complex species have created communication channels to rectify their sensory limitations to survive. Homo sapiens has widened their communication channels for faster throughput and started storing them as culture and carrying them through education. As a result we have created social truth.
Factual datas are the useful snapshot of the objective reality, a totem, a physical object can be observed with the sensors. Truth is an individual subject, an interpretation of the sensory data, a useful compromise.
The social truth is the useful compromise for the group by the group. The goal of the social truth is to survive as a group.
Physical Transcriptions of these social truth legitimise them.
We are tribal animal. We live in as a physical tribes and inside of hundreds of meta tribes in simulation which is the socio political data space we call it as the world.
Since we can’t access the objective reality reliably we look for social truth as the best guess blindly.
Institutions legitimise truths.
Fact driven institutions are more useful in the survival of the specie.
In other hand opinion driven institutions are not so useful for the species.
We do what we can get away with and exactly as expected within the context of our meta tribes.
We have two bodies
The biological one is like looking the earth from space.
And the political body is like the state.
The name you carry is the political body.
It transacts with the political states on the boundary less earth.
From the evolutionary perspective every biological entity has a basic feature which is homeostasis.
It’s the functioning sweet spot of that entity.
A control center read the sensory data to regulate itself to that state. By doing so it’s validate or update it’s prediction model.
In the process of becoming a complex organism it developed an extra layer of processing.
That’s our conscious mind.
And the control center remains as subconscious.
The subconscious collect the sensory data and regulate itself to stay functional.
Now when it stumble upon a novel environment it float the management to conscious mind to find the solution for homeostasis.
This conscious mind have one sensor which is language.
It works like a spiderweb.
As a spider creates it’s web it’s perception gets expand.
We are like spiders in a jungle.
We started creating these small webs at least 2/3 million years ago.
Our offspring stayed on it’s ancestral web reinforced it expanded it.
In time nearby webs became larger and connected with each other.
A common structural geometrical pattern emerges from this. This became the symbols which is the backbone of all language systems.
In time the forest becomes the mesh of web.
The superstructure is exactly the same but when we zoom in we can find different species of spiders are making their type of webs in between the super web.
Each spider try to senses the vibration of flies and
Try to catch it before others.
Every movement is telegraphic in the zone.
Every form of perceptions are just a different pitch of note traveling back and forth in the web superstructure.
There is a echo of older vibration pulsating through the web. Full of noise and self repeating hum.
That’s cultural history.
In the background there is the base hum in the infinite feedback loop.
Insignificant but ever present.
The sum of all the vibrations from the start.
super underrated channel, good stuff!
One of the few truly great channels. I can't wait to see more!
Yet another excellent video! Thanks Brit & colleagues
This is one of my favorite educational channels!
As always great content.An amalgamation of different concepts created coherently
Best channel i have seen in a while.
please share with others, UA-cam algorithm ignores this channel
Of course i will :)
Thank you so much. Im currently working on an AI myself thats more focused on the evolution and im working on in life memory that i will see if the AI will use to its advantage
love the idea...
i thought this video had 1M views atleast and released only recently, turns out it's been released in 2019 with low views... This is a high quality video! it deserves more attention!
Such a mind blowing video. The moss thing especially kind of blew my mind.
thrilled to hear it stay tuned for more (me too)
Best explanation of Intelligence
Learned so much more in 16 minutes. Awesome channel.
excellent!
Great job translating whats usually pretty dry philosophical and technological concepts into a compelling story.
I do think it's important to not judge historical approaches to learning models (e.g. expert machines) as somehow flawed or incorrect though.
Just like everything else, the space of computation ans mechanized calculation has changed drastically over the past 100 years, and that more rigorous well defined and logical approach to computation laid the foundations required for modern machine learning tools.
Building predictable and reliable machines that operate without intelligence or learning is the equivalent of constructing the environment in which higher order computation can occur.
Life wouldn't be possible without supernova and geological processes that operate without any form of intelligence, but operate predictably and give intelligence something to base itself within.
LLMs and Transformers are the higher order computational mechanisms that have the ability to learn because of the reliability and predictability of the more rudimentary attempts at creating machine intelligence that came before. Attempts that ended up creating an environment so predictable and reliable that those systems can form.
Wow love this, did you just find this video or did you already watch the LLM video I posted yesterday. If not i'd looove your thoughts.ArtoftheProblem
To me, the million dollar question (which was also asked by Church and Turing) is whether human brains are turing-equivalent or by some currently unknown process even surpasses turing machines. The church-turing thesis coined this concept 'effectively computable'; an algorithm is effectively computable if it can be carried out on a turing-equivalent machine, but in many ways our brains seem more powerful.
Lovely videos from you as always. If I might humbly suggest a future topic for this series: The challenges in machine learning when it comes to debugging or figuring out the individual "steps" that lead to the neural network's answer, and how that intertwines with moral philosophy when machine learning is used in applications for decision-making. Self-driving cars is one of many possible examples where this conundrum comes up. (Or I suppose it's not an issue iff you subscribe to Utilitarianism/Consequentialism ;) )
thanks for the notes/ideas. did you see part 3 yet? I'm currently thinking through part 4/5 and I'll definitely touch on sequential problems. history seems to be unfolding in real-time right now
We are not even Turing complete since we only have finite memory 👻 so the answer is no.
i have learnt so much about neural networks and evolution from these videos
thanks very much ❤️🥰❤️
thrilled to hear it, i hope to finally finish this series once I get my head around sequential networks :)
Very good video. I question the use of 'random' so frequently but that's flavor at the end of the day.
Very impressive writing. Lots of take aways. Thank you very much for your work.
appreciate the notes, glad you found this
A really nice video, once again ! Thanks for the effort you put into this, it's really useful !
Appreciate your feedback, I'm thrilled to hear it's useful to you
I'm going to use "arbitrary mental simulations" instead of "stories" now.
Awww, poor baby is unhappy with his life 😂😂😂
I wonder if our brains will ever evolve enough to communicate more accurately. Yes, we can mumble a few sounds to express something, but the interpretation is never exact. Can you imagine a computer that got a few 1’s & 0’s wrong, it would fail. Can you imagine if we could communicate exactly? I think it would be amazing.
1s and 0s are akin to yes and no.
Hence we ask politicians yes or no questions in attempt to attain exact answers.
McKenna noted that someone said about language, that it was invented in order to lie. It certainly enables a large fudge factor.
the most underrated chanel
thanks! please help me share it :)
You speak of AI but not yourself. Every video, you make the material so clear that even a child could learn it, AND THAT is intelligence, of the Organic kind!
I always have a child in mind when working on these, I sweat over each word. I'm glad you responded to this.
My pleasure, and thank you so much for the effort you put into these videos! Your channel maybe small, but we also only have 50k billionaires out of 7.8 billion folks ;) your channel is gold, those who know it's value are bound to sub! Keep it up my friend!
This channel is so well done.
A better way to conceptualize it is successful copies moving *downhill* not uphill. Gravity naturally pulls things down so it would make more sense, visually, to see the bacteria or data points, trending or sliding down into a hole rather than up onto a hill.
This was such a great video! Thanks a lot for the solid work.
thank you for the vote of confidence I appreciate it
Hey thanks for all the beautiful videos! What were your sources on the movement of bacteria looking for food?
Thanks glad you found this, I'm at airport so don't have sources on me, but please google it and let me know if you still need help finding
@@ArtOfTheProblem yeah i already googled it and found the chemotaxis of the bacteria, but i was just curious if you had a source that you found particularly useful for the subject. If so, let me know!
For background: i am researching predictive coding for my thesis and it seemed interesting that a bacteria already makes kind of a prediction about the location of food.
The BEST DEEP LEARNING CLEARLY EXPLANATION I HAVE EVER SEEN! 🎉 JL
thanks so much, did you see the whole series? my mother told me I still should make the 'for dummies' version
Thanks for the wonderful video! You have covered a LOT of topics in this relatively short video. I was hoping you would also include sleep as it is the least understood concept in scientific research and yet it dominates our daily lives.
thank you, so true I think about sleep a lot. I have the next 2 videos roughly planned out and perhaps I'll add a 3rd video to talk about unexplored ideas. stay tuned!
I have a naive and unsubstanciated feeling that this mold holds some secret that migt unlock general and scalable Q-Computers
i concur
Great video. I love how your videos inspire questions as well as give answers. Like what is the origin of intelligence if the first cells come equipped with cellular intelligence? My thinking being how do we explain a “muse” where do things like original ideas come from? If intelligence is compiled information over time? Also how does slow step by step evolution explain the absolute overwhelming complexity and length of the DNA chain? There hasn’t been enough time in the universe for it to work itself out by random chance, has there? And can we talk about the three problems with Darwin’s theory? The origin of DNA, the irreducible complexity of the cell, and the paucity of transitional species? I understand I may sound ignorant to some but that’s fine I’m here to learn. And the feedback you get from this channel is the best I’ve seen! Great job guys!
Thanks so much for sharing. appreciate it. I wonder if there is such as thing as 'original idea' or just 'new combination of ideas'
Damn youtube! this man is so underrated, would desire like a 10 million subs, for sure!
thanks :) this comment might be the one that breaks the camels back
New video is up on Evolution of Intelligence ua-cam.com/video/5EcQ1IcEMFQ/v-deo.html
Absolutely amazing brit! Thanks!
I really appreciate the feedback, i'm very proud of this video and will be posting the next one shortly
so clear
Your channel is pure gold, thank you for these inspiring videos!
appreciate this, working hard on next video now
Beautiful, just beautiful video.
so happy people are finding this!!!
congratulations bro@@ArtOfTheProblem
I love you man.
Incredible content! Please keep making things
Who ARE you. Your video describes exactly how I found you in the first place. I smelled around UA-cam randomly and then walked straight when I smelled your channel. How did I smell your channel? Totally random: I just happened to see a post you left SIX YEARS AGO on a CYCLIST'S YT channel (he's also a scientist I think but that's not why I was looking at his channel). Uncanny yet perfectly logical.
Your work really needs to be seen. I'm not a scientist but I am a teacher. I will share this and What Is Deep Learning? with my colleagues in September. Congratulations on achieving a fantastically succinct yet lucid synthesis of information.
wow I appreciate the kind words and the connections. Good to know it's even findable on UA-cam. Thank you for sharing my work it's the highest compliment
After watching this video my mind => KABOOM !!!!!
woo ! glad you found this series
So thinking abstractly is akin to just running a simulation without physical sensory inputs. Does this mean (perhaps in a very superficial way) that computers learn abstractly (e.g. AlphaGo)?
Regardless, fantastic video yet again. Keep up the good work!
I would say that's a byproduct of being able to separate senses into 'conceptual buckets', so yes it would. another example is the first neural network that just tried to learn 'square vs circle' in an image - the goal of generalization (abstraction) machines was this 'separation' of patterns to define a unique state.
What a beautiful work of Art!
thanks for the words of support, stick around as I'm working hard on the next few
Awesome as always. I'm looking forward to watch the rest!
very nice very nice
AMAZING presentation, as always! May I share something because I'm having goosebumps? ( Disclaimer: forgive the length of my rant, I'm blaming this Balinese coffee but it's just for casual reading for whenever, treat it as entertainment if you will - also CMIIW don't just take my word for everything I say here ).
11:14 corresponds NICELY to a recent 2024 breakthrough study regarding neurons and 12:55 NOTCH2NL suspiciously seems to as well - although I think the latter not as cause but instead an aftermath reaction.
I then also watched a standup comic deliver a punchline: "Now, animals aren't supposed to be wrong..." and everyone started to laugh. The comic unknowingly couldn't have been more in-line with the study while describing what we knew all along, which was why everyone laughed at that point.
According to the new findings: Both animals (mammals, vertebrates) and human have dendrites that were found to work in the same way... until the animals grow into adulthood - because human neuron development will remain stagnant. Animal synapses will mature into its final form where a lot of it become two-way neurons, receiving and transmitting can now go both ways, while human neurons will work the same way through life since birth.
My take on this (longform novel writing warning), perhaps this is _the_ dead-set mechanics post an animal's essential learning chapter where they became predictable and are "never wrong", becoming one with nature and "majestic" to see, as we usually describe them.
I reckon long after we're done being a dead-set serious adult tree rat ourselves, evolving just like the rest of majestic life and mutate some limbs or two, somewhere later maybe before Lucy our neurons lost the ability to conclude its growth - never maturing like it was before, staying juvenile, explorative, as an error-loving entity - just like those fun loving and silly baby animals, pups and cubs and those annoying persistent kittens and whatnot did before their neuron maturity.
We never grew up, because our neurons don't finalize anymore.
If that is so then I think the consequence is an ever hungry juvenile brain for life.
By the way, people call this INIT_FINALIZE ERROR as: Intelligence - though typing from my state-of-the-art laptop to reach you from across the globe via a satellite our species managed to put in space I failed to confirm this claim to be agreed with with other animals sharing the planet. So for now we'll just lay claim to this predicate in solo.
If I have to pull something out of my I'd imagine somewhere along the great ape family tree a group had offspring that weren't born complete and stays as pondering children wreaking havoc with their juvenile why's and what-if's - yet managed to out-populate their "neurotypical" peers - perhaps because their mating standards became mentally exclusive as these rowdy "misfits" find each other more interesting - or more forgivingly understanding - and do things the same novel way. But whatever actually really happened, an obviously significant energy re-routing happened, probably even left a detour marker in time within our evolutionary slope, when a group floored the brake pedals to finally stop mutating in form because they became hungry for energy in a different manner, as Dr. V.S. Ramachandran once said (paraphrased, he didn't say it exactly this way): "Polar bears developed massive fat tissues, thick skin and white fur to survive in the cold while we develop better spears, knives and sewing kit to eat the polar bear and use its fur to do the same thing"
...so we became the title holder of "Most Dangerous Animal In The World" while at the same time won't last a fighting round to a murderous chimp in a ring.
Still baffled though, that mother nature would let a loss in function to become a winning survival strategy such as ours that even left poor old maturing-dendrite grampa in the dust - or is it not? That this was just another type of a common "unsustainable" glitch that scientists are yet to notice in other past or present species?
What I meant was that a thing could be made to work but doesn't mean that it's right, a mis-procedure posing own risks in time. Hominids peculiarly have no stopping boundaries in everything including populating a whole planet. When we battle, other species could die along, isn't it weirdly unique? We compete, conquer, dissolve each other into a single species group (which I guess viruses will love) and aside from "one man standing", our other counter-intuitive popular concept to survival is "one man's loss is another man's gain", we may also become either parasitic or mutualistic towards nature by choice and not out of necessity - heck we're currently even planning ways to leave this planet altogether and start a new civilization elsewhere.
Our solely claimed Special Ability could actually well be nothing but a side-product of a Special Disability turned pragmatic, artificial method that happened to work so far on a planet that constantly tries to kill us, so we kept on saying... or rather, due to unspecialized mutation i.e. quitting the institution at Great Ape, was the whole reason why we felt that way while the rest of the animal kingdom might not agree - I bet if animals could talk and we asked what made them survive nature they won't answer the same way we do because we'd be asking nature itself, they'd probably say "Huh?" while we'll just have to settle with "Because we'll do our best" - and that's because we have no choice but to use something else to perform it, almost as if we're not natural anymore, even though once we definitely were....
...which leads me to end this silly rant of mine, in your style (with a question), regarding the future unknowns and unforeseen:
If we _were_ an honest kink in the fabric, a loose thread - then for how long can we get away with it?
Love the goosebumps! FYI consider supporting future content via. www.patreon.com/artoftheproblem - thanks again
@@ArtOfTheProblem Will definitely do!
new video is out would love if you could help me share it around, I only have 24 hours left for the algo to catch it: ua-cam.com/video/PvDaPeQjxOE/v-deo.html
damn, that was good. very clear way of explaining these learning layers .
appreciate the feedback on this video. It's very much a 'thought in progress" stay tuned for next video
Amazing channel, great job !
thanks, hope you enjoy this latest series, two more videos on the way in it.
I have a question. I just saw your video explaining how the network works by combining the information together and subtracting the original information to get the other information, and I wondered, if you don't take away the information that is subtracted, won't it be possible to obtain information that you never searched for? If this is possible, wouldn't it be possible to obtain personal information without doing the complex procedure of hacking through the security?
I love your videos, it's so easy to understand C.S. with these videos. Keep up the good work!
hey, are you referring to the network coding video? You should ask your question over there and I can answer. thanks!
This is a really good video on this topic only 18k views? crazy
i know...
There is an interesting (popular) book called "The master algorithm" by Pedro Domingos, where he presents the 5 schools of thought in (weak) AI (symbolists, evolutionists, connectionists, bayesians and analogizers), their primary algorithms and tries to integrate them into a future joint AI framework. Somewhat similar to what you did here, but less from a biological, and more from a computational view. Obviously he also ends up with a bottom up hierarchy. He's a Prof for AI at the University of Washington.
Yes this book was part of my research. I really enjoyed the book but I felt that the final chapter where he tried to pull things together in that strange analogy fell flat. Do you know what I mean?
Art of the Problem
Yeah, I guess that's because nobody has been able to pull it of to unify all of them into a single framework, so of course he is kind of fantasizing. Regarding your series: especially Information theory: I'd suggest you could go more in depth. Don't be afraid of formulas.
BTW, there is a great book called "Information theory - A tutorial introduction" by James V. Stone. In contrast to most Information theory books, it is quite mathematical, without being mathematical, so its actually quite readable and also covers e.g. the non-discrete values.
How do single celled bacteria remember? Do we know the mechanism?
Do you plan on continuing this series? Love it so far.
Yes I will publish part three later this month, stay tuned
Awesome
Amazing! 🙌🏼
The video content basically unifies in a single theoretical framework organic and inorganic intelligence forms.
This was my hope while researching this, I felt slightly crazy making it because I went a bit 'off book' while writing it...I was waiting for the backlash....
@@ArtOfTheProblem Absolutely fantastic, congratulations.
@@bardes18 I know, I find it's best to use simple games like tic tac toe, then chess as one avenue. Lately though, I find that if you just explain to people that we are building a 'new computer' which is modelled off 'our brains wiring', and letting it 'experience the world'. they seem to 'get it' without knowing the details...the analogy is, we literally copied the brain with wires that seems to help it click for some. others i've found need to know that 'these networks could also just be copper wires' to understand that this isn't 'magic' but just very simple engineering. in the next video I hope to clarify this by building a tiny physical neural network, then expanding it, and walking through the first experiments...so you see exactly where the 'magic' occurs (which is when we chain together layers, and get an exponential payoff in functional complexity...the whole, you could use 10 billion neurons in one layer, or 30 neurons in 3 layers and do the same thing). and then i hope to connect all of that to the math theory where "thing = unique shape in space (manifold)", and the network is actually morphing the space to isolate concepts (separating shapes)
Nice arbitrary mental simulation
I'm sitting here thinking about the inner workings of abstract thinking - how communication between one mind and another lets you imagine one person's ideas and fit them into your own and actually learn new things that way, because being told something new has to be understood in some abstracted way first in order to learn from it, otherwise it's just raw perceptions? Did I get that right? Also, kinda trippy that this same process is going on while trying to think of it. That adds a whole new level of complexity I feel like.
Exactly, did you know you are quoteing my information theory series??? please check it out
@@ArtOfTheProblem No I just conjured that up on my own haha. But I'm glad to hear that because it makes me feel like there's some definite truth to it. I'll definitely be checking that out now. I just discovered your channel last night and it has been great
premium content right here
Immanuel Kant would be proud. Somebody had to nail synthetic a priori eventually.
Wow, thanks!!
glad you found this series, how did you like the rest? more to come
How can one create such masterpieces for so long, and still not break through.
In the meantime, those click bait misinformation "cash cow" channels are generating huge amount of views.
What a sad world we live in
thanks for saying it, i appreciate it. I "think" it's very irregular output. I gave up on the breakout goal a long time ago but was surprised to see my last video did really really well. i wonder if I can repeat that on another one or two this year. working on RL now but it's going to take a while
@@ArtOfTheProblem it could be, but in a way, you're very consistent at generating factual videos. Googles algorithms could seriously steer humanity towards a brighter path, instead, it mostly shows me clickbaity videos.
If I can give you a tip, I find a lot of videos through shorts, but shorts are a though business where quick dopamine bursts are all the watchers are after. Your shorts are super super informative, but not so good at hooking the watcher. Perhaps you could try starting with a flashy question first: "how do computers even learn?!". I know it's terrible, but it could work :/
@@insevanhouts I will try this !
would love if you could help share my newest video: ua-cam.com/video/5EcQ1IcEMFQ/v-deo.html
Pure Gold 🥇
But, background music is unnecessary and super annoying.
wicked as usual ;)
You are a great teacher ☺️
thanks for stopping by
Amazing video just a note though at 6:17 you state that the shock ( aversive stimulus) acts to negatively reinforce a previous unconditioned stimulus. This is incorrect as applying any consequence that decreases a behaviour will be a punishment. Where as negative reinforcement increases the frequency or likelihood of a behaviour occurring by removing an aversive stimulus. A slight difference but a important one as a negative reinforcement cannot decrease a behaviour or else it is not a reinforcement.
great catch
Calling it modern is somewhat biased I think. It gives the impression that the top down model is obselete, and it is not. It is also important to note that higher intelligence is a lot more complicated than simple machine learning. It's not just some quantum annealing process to determine the shortest path, it's also what figured out what that process is. The human brain is subject to the influence of all forces of nature, and like the Universe itself it's immeasurable by a closed system of functions. When we think about the theoretical God machine for example, to scan the entire Universe and simulate the future which will absolutely take place is an impossibility. Even if you knew the exact position of all particles and simply applied thermodynamic laws, you would still have to include the machine itself in that calculation and such a machine cannot store its own state so it will never be able to absolutely predict a future which is the essence of the Heisenberg principle. Also, as you noted, human brain may not be the best approach in all cases as mass extinction events have proven to us. If a dog listened to low quality digitally generated sound it would sound a lot more like what 8-bit music sounds like to us and that's due to resolution. That simple difference in the brain may give a dog more "intelligence" given a situation where hearing discrete differences in sound would provide an advantage so as you can see all of life is a historical record if not the Universe which is the essence of non-locality. Other than the human ego biases great video. I will give it a plus square root of 2 😀
Amazing video
thank you! stay tuned for follow up
just posted new video on RL ua-cam.com/video/Dov68JsIC4g/v-deo.html
I really need to obey the algorithms here and remember to always clicking on the thumbs up icon. And always add a comment (increase social engagement ranking) to all the videos on all channels I love. This content should be push to far more potential eyeballs by Alphabet Inc than they do. The current system is, by its clickbait nature, designed to eventually degenerate into something like what was portrayed in Idiocracy, such a prophetic film. I much obey the algorithms :)
I gave up on the algo a long time ago, when it wanted weekly videos which isn't possible no matter how many resources I threw at the problem. It's like BBQ, low and slow is the only way. thank you for your support it means a lot.
Wow thanks for this video
appreciate your kind words, stay tuned for much more
great
Teach octopuses to write stuff down... They are held back by short life spans, only.
Subscribed.
welcome to the family!
13:07 Its big brain time! awesome video
loved this
appreciate the feedback, check out my latest vid onRL
new video! ua-cam.com/video/PvDaPeQjxOE/v-deo.html
Hey! If you can help share my new video around any of your networks today it might catch fire and would help me support the channel. I appreciate your help! ua-cam.com/video/PvDaPeQjxOE/v-deo.html
🙏
stay tuned for more!
thank you.
glad you enjoyed stay tuned for more
Next video?Eagerly waiting..
writing next 2 as we speak, hope in the next month I get it out
@@ArtOfTheProblem 😍
they would have a left and right our nose has a central point and ears have two,it most have nose or body that smells or taste.Also when their in 3d its more difficult.Dogs do something like your explaining like its math,they will sniff then jump and sniff,then they see or lad straight on it instinctual
And we ALL know humans who go on and on!!!
:)
This is deep
thanks! yes I try to go as deep as possible
genius video. please keep it up
thank you, let me know what you think of the rest of the series
@@ArtOfTheProblem Il be watching them all