How Does Each Brain Make A Mind? Adaptive Resonance Theory (ART) & Consciousness | Stephen Grossberg
What this covers
Stephen Grossberg is Professor Emeritus of Mathematics & Statistics, Psychological & Brain Sciences, and Biomedical Engineering at Boston University. He is Wang Professor of Cognitive and Neural Systems & Director of the Center for Adaptive Systems. He is a Cognitive Scientist, Theoretical and Computational Psychologist, Neuroscientist, Mathematician, Biomedical Engineer, and Neuromorphic Technologist. He has published 18 books or journal special issues, over 560 research articles, 7 patents and 100 000+ citations. He has been recognised for the past 50 years as the most important pioneer and current research leader who explains how our brains make our minds. Grossberg is often called the Einstein of the Mind.
TIMESTAMPS: 0:00 Introduction 0:32 Steve's groundbreaking work on the Stability-Plasticity Dilemma & Adaptive Resonance Theory (ART) 10:58 Steve is "The Newton of the Brain" or "The Einstein of the Mind" 13:53 Competitive Learning & Catastrophic Forgetting (Learned Expectations) 21:31 ART explains CLEARS! (Consciousness, Learning, Expectation, Attention, Resonance & Synchrony) 30:17 ART's Explanatory Power & Predictive Success 37:57 ART is now a pioneering field in science & mathematics 41:51 Cognitive Emotional-Motor (CogEM) Model & Adaptive Resonances 45:57 Explaining & Predicting Mental Disorder with ART (ADHD, Schizophrenia, Alzheimer's, Autism) 54:38 Surface-Shroud Resonance (Difference between Conscious Seeing & Action) 1:07:18 Autonomous Adaptive Intelligence 1:15:36 "Conscious Mind, Resonant Brain: How Each Brain Makes A Mind" Steve's Magnum Opus 1:33:17 ART on Creativity & Religion 1:44:26 Recent developments in ART 1:49:41 Steve's message to future scientists 1:57:58 Steve's Impressive Legacy 2:15:10 Conclusion
EPISODE LINKS: - Steve's Round 2: https://youtu.be/gpa0beB18vk - Steve's Website: https://sites.bu.edu/steveg/ - Steve's Books: https://tinyurl.com/2jjvvbcs - Steve's Publications: https://tinyurl.com/4mcr4pbk
CONNECT: - Website: https://tevinnaidu.com/ - Podcast: https://podcasters.spotify.com/pod/show/drtevinnaidu - Twitter: https://twitter.com/drtevinnaidu/ - Facebook: https://www.facebook.com/drtevinnaidu - Instagram: https://www.instagram.com/drtevinnaidu/ - LinkedIn: https://www.linkedin.com/in/drtevinnaidu/
=============================
Disclaimer: The information provided on this channel is for educational purposes only. The content is shared in the spirit of open discourse and does not constitute, nor does it substitute, professional or medical advice. We do not accept any liability for any loss or damage incurred from you acting or not acting as a result of listening/watching any of our contents. You acknowledge that you use the information provided at your own risk. Listeners/viewers are advised to conduct their own research and consult with their own experts in the respective fields.
#StephenGrossberg #Consciousness #MindBodyProblem
Source description (no synthesized summary yet).
Grossberg argues that Adaptive Resonance Theory (ART) provides a unified mechanistic explanation of how consciousness, learning, attention, and prediction emerge from brain dynamics, solving the mind-body problem through computational models of neural self-organization rather than through philosophical speculation or statistical pattern-matching.
- ART's resonance mechanisms link consciousness, learning, attention, and synchrony through a single causal architecture that operates autonomously in real time
- The theory emerges from analyzing hundreds of psychological experiments to discover design principles, not from assumptions or mathematical convenience
- ART explains both normal cognition and mental disorders (autism, ADHD, schizophrenia, Alzheimer's) by showing how imbalanced opponent processes and vigilance control produce specific symptoms
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Adaptive resonance occurs when both bottom-up adaptive filters and top-down expectations simultaneously act together, creating a positive feedback loop with excitatory signals between bottom-up and top-down pathways; when patterns match well enough to cause active cells to synchronize, the positive feedback sustains synchronized activity at enhanced levels, which is sufficient to trigger selective learning.
“a resonance can develop so when all it resonances occurs when they're excitatory feedback signals between the bottom up and the top down Pathways between two or more brain regions and when these patterns match their signals well enough to cause the active cells to synchronize you know a lot of brain rhythms synchronized and because of the positive feedback loop they s they sustain this synchronized activity at an enhanced activity level and that's enough to trigger learning but in a very selective fashion”
Brain evolution needs to achieve behavioral success, so without brain mechanisms one cannot understand psychological functions including learning, but without psychological functions the brain mechanisms have no behavioral meaning—therefore understanding any psychological process requires simultaneous analysis of both brain mechanisms and behavioral function in real time.
“brain Evolution needs to achieve behavioral success so without brain mechanisms one can't understand psychological functions including learning but without the psychological functions the brain mechanisms have no behavioral meaning so to me I always felt the need since I was always thinking in real time about how does this kind of learning or other Human Experience emerge Moment by moment in real time I had to think about what's going on in there to enable it”
Superstitious behaviors arise from accidental correlations between events; if you enter a dining room only when food is served, you might associate the wallpaper with food appearing, even though the wallpaper doesn't cause food; many superstitious rituals and primitive medical practices arose throughout history due to low-probability coincidental correlations.
“few of us would say that the wallpaper on a dining room wall causes food to appear on the dining room table even though it's present whenever food is served on the other hand if you went into the dining room only when food was served you might very well associate the distinctive wallpaper in the dining room with eating”
The mathematical neural models the speaker developed resembled biological neural networks, validating that learning to think in real time about neural dynamics produces models that match biological reality—this is an art that must be learned rather than a mechanical procedure.
“the big surprise to me when I was a boy of 17. is that these models look like you know they they help me operate it and the model themselves looked like biological neural networks and the bottom line is learn how to think in real time that's an art”
Cognitive and emotional processes interact to enable choice of actions that realize valued goals; art has no concept of values—emotions and feelings are necessary to direct thoughts toward worthwhile goals, integrating cognition, emotion, and motor systems into a unified model.
“you need emotions feelings to direct your thoughts to valued goals worthwhile goals and so at this point I like distinguishing six different kind of adaptive resonances that I've studied um uh they support different aspects of Consciousness you were saying that's what's triggering this thought with different functions in different parts of our brains”
Coordinate transformation is necessary for visually guided action because we see objects in head-centered coordinates but reach toward them in body-centered coordinates, and these relationships continually change throughout life as body proportions change.
“you have to solve the number of coordinate change problems to do it because you're looking in head Center coordinates you're reaching in Body Center coordinates and with colleagues like Dan Bullock and the greed and Frank Gunther we spend a lot of work showing how we self-organize these coordinate representations in bodies whose uh relationships between eyes and neck and shoulder are continually changing during uh Our Lives”
I introduced the field of neural networks in 1957 when I was 17 years old as a college freshman taking introductory psychology at Dartmouth College, inspired by basic facts about how humans learn that led me to derive my first neural models of how brains make minds.
“my life's work began in 1957 before you were born that I was 17 years old and I was a college freshman taking introductory psychology at Dartmouth college and it was in that year that I introduced the field of neural networks”
Most of my work emerged from analyzing self-organization in both development and learning, particularly solving the stability-plasticity dilemma—how we learn quickly without catastrophic forgetting while also maintaining stable memories—which I addressed through Adaptive Resonance Theory (ART) introduced in 1976.
“most of my work emerged from an analysis of self-organization both in development and learning notably how we can learn quickly without suffering from catastrophic forgetting otherwise expressed how we can learn quickly without forgetting Justice quickly which of course was a topic very much on the mind of a young student and learning quickly without catastrophic forgetting I'd come to call solving the stability plasticity dilemma where the plasticity is the learning and the stability is the fact that memories are buffered so you don't have catastrophic forgetting and that's how I was led to introduce adaptive resonance Theory or art or just say art in 1976 to solve the problem”
To explain a hundred facts with one model requires balancing and harmoniously resolving multiple constraints—this is the real challenge in theory-building, not ad-hoc explanations of individual observations.
“try to explain 100 facts in one way you know that's where the challenge is it's trying to see how well the constraints can be harmoniously resolved”
Do good works and try to make the world better in whatever way you can—we are only here for a short time, and one form of immortality is having positive thoughts about you after you're gone; a troubled world needs talented people contributing in every possible way.
“try to do some good in the world you know we're only here for a short time um try to make it so you know one concept of immortality is people will have positive thoughts about you being here for the brief time you were so do good works you know that's that's wonderful advice Diva I mean you you've got an age index of around 130”
The power of the thought experiment is that if you accept the simple hypotheses (familiar facts of life), you must either accept the conclusions or reject the scientific method—showing where errors lie in the thought experiment becomes a way to test whether one believes in logical reasoning.
“someone says oh I don't believe any of this nonsense and then you say Okay read the thought experiment show me where he made a mistake because if you can then if you believe this simple hypotheses which we all know in that area die they're Facts of Life then you have to either accept the conclusion or throw away your belief in the scientific method”
Future scientists and philosophers should find a problem they are passionate about—one they are willing to spend hundreds of hours learning and thinking about—because real progress requires sustained commitment; not everyone should become a theorist or experimentalist, and alternative careers (teaching, other professions) are also valuable ways to contribute to society.
“find the problem you're passionate about one that you're willing to spend the hundreds of hours that you're going to need to learn and think about it anything if you're going to make any progress at all and not everyone is going to be passionate about a problem on mind and brain or indeed and science well then don't become a theorist don't become an experimentalist you know for example if you still like reading stuff in science or other areas one Noble profession is teaching which is very satisfying profession”
Grossberg's key fundamental contributions include introducing the paradigm of non-linear systems with differential equations to link brain to mind through neural networks in 1957; introducing shunting on-center surround networks solving the noise-saturation dilemma; introducing basic equations for neural modules and micro circuits used throughout the brain.
“I introduced the Paradigm of non-linear systems with differential equations to link brain to mind through neural networks in 1957 I introduced the somewhat larger number of modules on micro circuits uh this is all on my Wikipedia page”
Doing science requires maintaining pure emotions and emotions and intellectual integrity to avoid letting petty feelings drive non-intellectually-valid work; a scientist's character shows in how they respond when evidence contradicts them or when their work is questioned.
“a lot of doing science is having good values I I felt that I have to keep my emotions pure so I don't do things that are not intellectually valid because of you know Petty emotions”
His legacy consists of the equations, modules, and architectures he introduced; he combines specialized versions of micro circuits and equations into modal architectures for different modalities; this represents the integrated theoretical framework that is his lasting contribution.
“that is the theory that's how it's organized and and yeah I've gotten partial credit for it and I am grateful for it and if I if you work through that you'll never never keep going because the whole point is to be ahead of the game”
ART has evolved into multiple varieties and applications—as people apply ART in technology they want the most efficient algorithm for their problem, leading to specialized versions of ART for different engineering and AI applications.
“well just with art there are multiple varieties quote you know when people are applying or in technology um they want the most efficient algorithm to solve a certain class of problems and typically just as in our brain before you get in for temporal cortex and prefrontal cortex and other cognitive areas there's a huge amount of pre-processing visual pre-processing”
John Kennedy, Einstein's last assistant and co-inventor of computer time-sharing systems, was a mentor who understood solitude and loneliness from working with Einstein; taking his course on 'A Philosopher Looks at Science' was transformative because it taught me the importance of philosophical grounding for scientific work.
“luckily I had two wonderful mentors one is John Kennedy have you ever heard of John Kennedy well he was Albert Einstein's last assistant oh wow because he co-invented the basic computer language he co-invented computed time sharing systems so he was very into computing and having been so close to Einstein he knew what it meant to be a loner and he also wrote a wonderful book called a philosopher looks at science and I was lucky to take a course that read out of that book when I was a sophomore”
Mental health disorders arise when cognitive-emotional and attentional-orienting processes get imbalanced or damaged: symptoms like those in Alzheimer's, autism, amnesia, and PTSD emerge from specific imbalances in these mechanisms, and studying normal brain theories reveals what goes wrong to cause disorders.
“well first I think it's really useful to have brain theories that explain a lot of facts about normal or typical Behavior because it's only when such processes get imbalanced or damaged in certain ways that the behavior is no longer typical and if it gets imbalanced enough to lead to symptoms that people find not very sociable then we'll call it a mental disorder so I was led through the back door just like with Consciousness into mental disorders by noticing hey if the following processes and cognitive and cognitive emotional Dynamics get imbalanced symptoms of Alzheimer's or Autism or Amnesia or post-traumatic stress disorder Etc pop out”
Discovering underlying principles from large psychological datasets requires a speculative leap—a gift that cannot be algorithmic but must be learned by studying with someone who does it well, similar to how Einstein found principles underlying physics.
“it was a speculative leap if you live in hundreds of experiments long enough you can begin to see what the underlying principles and mechanisms are that generated those behaviors as emergent properties it's a gift it's a leap there is no algorithm for it and the way to learn it if you're lucky is to study with someone who does it well”
The speaker's 66 years of foundational work have proven very useful for diverse recent projects (music, art, meaning), showing the long-term payoff of deep foundation-building versus pursuing fashionable topics.
“that means everything that I'm doing I have to do with the foundation in place that's been built over 66 years so far the foundation has been very useful like for the music work the artwork the um meaning work”
Don Bunch at Missouri University of Science and Technology was inspired by ART as a young student and has continued to work with related models; Robert Heck Nielsen was inspired by the speaker's work during his PhD and founded HNC Corporation (later Heck Nielsen Neural Computer) focused on bringing neural networks to technology.
“one colleague a younger colleague but now he's a very senior guy in his own right he's Don bunch at Missouri Science and Technology... Robert heck Nielsen was Robin um used to work in a company in California and he would fly to have lunch with me when I was still at MIT and we talked about neural networks and he found that a company which was originally pulled heck Nielsen neura computer”
A key transition in the speaker's life was learning to stop competing to rank higher and instead learn out of love—this shift enabled him to make discoveries as a freshman, and he attributes this psychological transformation to good mentors.
“I said on top of that you were always first in your class I mean basically yeah so so anyway so I knew what I had to stop learning to compete which was a survival skill and start learning out of love and my seat and my freshman year I made my discoveries and I took camery's classes sophomore and he wrote that I wrote the best essays he had ever written he had ever read without qualification”
Individual neurons can respond selectively to certain combinations of features, and when they respond their activity (firing) creates what the speaker calls a short-term memory trace (STM trace), realized by a change in membrane potential, which is transient—the cell loses activity if nothing else happens.
“individual neurons or nerve cells can respond selectively to certain combinations of features and when they respond their activity often increases they're firing and such an activity I like what's it called a short-term memory trace or STM Trace it's realized in this by a change in membrane potential in the main reasons for short-term memory Trace is you know the it get hey can I get this in that it it is hard to do in the picture because of the angle the cell gets active for a little while and then it loses its activity if nothing else is happening”
All neural models I've developed can operate autonomously and learn either unsupervised (learning from inputs alone) or supervised (learning with feedback about predictive success); perceptual and cognitive processes are designed to be general-purpose—they can respond adaptively to any input pattern in their domain and to changing environments.
“all the neural models that I've discovered and developed and Illustrated operate autonomously and the ones that learn can learn unsupervised just in response to inputs from the world or supervised where you can um the supervision is provided by feedback about predictive success in the world”
Despite its complexity, the book is written to be accessible, with friends who are a rabbi, pastor, visual artist, gallery owner, social worker, and lawyer having enjoyed reading parts of it—readers can jump to individual chapters independently since each is written to be readable without reading the whole book.
“I have friends who are a rabbi a pastor a visual artist a gallery owner a social worker a lawyer who know no science who've enjoyed reading parts of it and I say parts of it because I know it's a long book and everyone's busy and if you read the introduction and preface you can then jump to any chapter you like because I wrote the chapters to be readable independently of each other”
Emergent properties of brain behavior cannot be fully understood using purely experimental approaches even with a hundred electrodes in a behaving animal—there is an explanatory gap between data and understanding that requires mathematical models to bridge.
“the emergent property and behavior there's an explanatory Gap and for that you need models”
Perception is general-purpose—you see or hear in response to both familiar and unfamiliar events—while recognition requires prior learning and only responds to familiar events: these are fundamentally different processes and must never be conflated.
“I already said perceptions general purpose you have it in response to both familiar and unfamiliar events but recognition is only to familiar things you've already learned about someone must never conflate them”
Opponent processes realized through gated dipole circuits solve multiple dilemmas including arousal regulation: two opponent channels compete and are gated by transmitters (medium-term memories) that habittuate; when a fearful stimulus triggers the on-channel and then shuts off, habituation means the off-channel rebounds, creating relief—the balance between arousal, habituation, and rebound explains emotional dynamics and disorders.
“a theme in a lot of the work not the only design theme is one of opponent processes which I model using uh ubiquitously occurring circuit that I call a gated dipole um dipole their opponent on and off cells that are competing and the gates the transmitters chemical transmitters these are like the medium term memories I earlier mentioned that can habituate if you use them”
He grew up in New York when Jewish quotas restricted college admission, forcing outstanding students to achieve exceptional grades and test scores to access education; this competitive environment taught him about academic rigor but also drove his early desire to escape New York and pursue intellectual freedom.
“I grew up in New York at a time when um it was a very hard for Jewish kids to get in college it was a Jewish quota so we worked super hard to get good enough scores on General tests and grades that we might get a scholarship to go to a good school”
Bayes rule is mathematically valid for statistical estimation but is merely a tautology stating probability relationships—it says nothing about the world without additional assumptions, and unlike physics or chemistry, it lacks heuristics for which assumptions to add, leading researchers to 'throw in the kitchen sink' trying to fit it to data, which is sad and represents trying to get something for nothing.
“the Bayes rule it's the probability of two events A and B written in two ways the probability of a given a times the probability of a or the probability of a given B times the probability of B it's just a tautology take that equality between those two ways of probability a given B times probability equal probability of being given a times probability they and divide let's say both sides by probability a day and maximize that's the base rule it's a tautology but it's used in statistical estimation some of my good friends are good statisticians who've used it but it says nothing about the world”
Too little arousal causes under-aroused depression where people are hypersensitive to stimuli; too much arousal causes over-aroused conditions where people are hyposensitive and insensitive to input—conditions like ADHD show hypersensitivity (under-aroused) while schizophrenia shows hyposensitivity (over-aroused), both representing failure to achieve proper arousal balance.
“and then if you have too little arousal you're you have a kind of under aroused depression and uh and then in response to a bigger than normal input you can be hypersensitive that's the paradox and you can get that in ADHD for example um and if you're over aroused then you always have enough arousal but you're very insensitive you're hyposensitive 10 parts you could get that in schizophrenia for example very insensitive”
Top-down modulatory on-center signals can sensitize and modulate neurons but cannot fire them to suprathreshold levels without support from a sufficiently well-matched bottom-up input pattern or additional gain control; this allows volitional modulation of neural activity (as in visual imagery through basal ganglia) without hallucination.
“the modulatory on Center can sensitize modulate or prime the cells to which IT projects but it can't fire them to Super threshold levels without support from a match that's sufficiently well-matched bottom-up input pattern unless you use additional gain control”
Unexpected visual cues that mismatch ongoing processing activate the orienting system, producing a novelty reaction and burst of non-specific arousal that drives search for better matching categories; if a cue is familiar, reset occurs quickly in one step; if truly novel, the system creates a new uncommitted population to begin learning.
“in unexpected cue will mismatch ongoing processing if there's a big it's enough mismatch it'll activate the orienting system which is either depending on the part of the brain and or hippocampal region or a non-specific calamic region uh and then you'll get a burst of non-specific arousal a novelty reaction you know and that will drive a surge for a better matching category”
I introduced short-term memory traces, long-term memory traces (adaptive weights) for learning and memory, and medium-term memory (activity-dependent habituation) where cells get tired if you find them too long—these are three fundamental temporal scales of memory that operate in neural networks.
“I introduced selectivations or short-term memory traces I introduced long-term memory traces or adaptive weights for learning memory and I also introduce what I like to call medium-term memory which is activity dependent habituation in a word if you find them too long cells can get tired”
Richard Atkinson, who became president of the National Academy of Sciences, and John Kennedy both praised the speaker's work as having potential to save mathematical psychology, providing credibility during career transitions.
“Richard Atkinson who then went on to be become president of the National Academy of Sciences who is a professor of psychology said oh the other one Kennedy said is going to save methodical psychology so I always had someone who was Ed cots wrote to MIT that I was the most creative person you'd ever met”
Deep learning and backpropagation represent semi-classical steepest descent optimization and are not the revolution; the revolution is in autonomy achieved through the canonical cortical circuit properly embodied in VLSI chips.
“because they're chips they can go in everything from algorithms and machines to mobile robots because they're light it'll take a big big effort and Google seems to be fascinated by Deep learning but if Google wants to make trillions of dollars if they do it right this is where the revolution is the revolution is in autonomy not in semi-classical steepest descent”
Surface-shroud resonances explain conscious seeing of visual objects and scenes: the retina has a blind spot where the optic nerve forms, but consciousness is not aware of it because cortical boundary completion and surface filling processes complete the retinal image at multiple processing stages through hierarchical resolution of uncertainty.
“and you complete the retinal image of the blind spot higher up in our little cortex by multiple stages of what I call boundary completion and surface filling in because I've shown that 3D boundaries and surfaces properly understood are the functional units of conscious vision”
Every animal or organism that can autonomously solve the stability-plasticity dilemma has a form of consciousness—consciousness is not unique to humans but is a fundamental property of any adaptive system that can learn and predict without catastrophic forgetting.
“I make for granted that every animal or other organism that can autonomously solve this stability plasticity dilemma has a form of consciousness”
The brain completes partial boundary and surface representations through hierarchical resolution of uncertainty, applying principles of uncertainty, complementarity, and resonance similar to those in quantum physics—raising the question of whether brain principles are related to physical principles with which our ancestors' brains evolved.
“their principles of uncertainty complementarity and resonance in our brains this is one of the certainly principles and if you think about physics lots of physical theories notably quantum theory have principles of uncertainty complementarity and resonance and so you are fascinated by the question of well how are our principles related if at all to the principles if the physical world with which are our ancestors have ceaselessly interacted while our brains were evolving”
Pre-attentive grouping like boundary formation occurs automatically before you pay attention to it and is the difference between perception and recognition; pre-attentive grouping is its own attentional prime due to how horizontal and feedback interactions work within laminar circuits.
“a pre-attentive grouping like a boundary which can form automatically even before you pay attention to it it's the difference between perception and recognition pre-attentive grouping is its own attentional Prime due to the way in which horizontal and feedback interact within the lamina circuits”
The paradigm of laminar computing shows that multiple brain regions can interact seamlessly because they all use variations of a single canonical cortical circuit for different modalities (vision, audition, language, cognition), suggesting the cerebral cortex achieves all higher-order biological intelligence through repeated application of the same fundamental architecture.
“what I call the Paradigm of laminar computing which clarifies have multiple parts of the brain can interact seamlessly together because they're designed as a variations of a single canonical cortical circuit even though they may be doing Vision recognition speech language uh just be advice you know fine-tuning a canonical circuit”
Read the published literature—in the age of Google there is no excuse for rediscovering the wheel or not searching out key papers; future theorists should study my work and that of collaborators, particularly the 2021 book 'Conscious Mind, Resonant Brain' and 500+ archival papers on my website as resources for understanding current theories of how brains make minds.
“read the published literature in the age of Google there's no excuse for trying to ReDiscover the wheel or not to put to find a point on it you don't want to waste a lot of your life just plagiarizing because it won't really help you in the long run and if you want to understand the latest theoretical Concepts and models about how brain makes mind and certainly include in your reading work that my colleagues have done and I wrote the magnum opus conscious mind resonant brain how each brain makes my”
I developed ART's predictive framework starting in 1976, well before widespread interest in Bayes rule, and ART is a predictive theory—but unlike Bayesian approaches, ART has foundational hypotheses that are already explanatory and predictive, representing 'the gift that keeps on giving' rather than trying to get something for nothing.
“remember I did Art which is a predictive Theory starting in 76 and all of the fed like interest in the Bayes rule was much later”
I did self-organizing systems work before Christopher von der Malsburg, who received government promotion in Finland despite ignoring my earlier published results and my written critique presented at a conference—his failure to acknowledge or engage with prior work represents poor ethics, and he doesn't deserve credit for claiming novelty on work others developed first.
“like like today of Oklahoma yeah who you know people credit with self-organizing that well no I did it before him yes and so did Christopher Vander Mossberg and I worked on in the mid 70s in Taylor started writing about in 82 and 84 and I was I was the session chair where he presented his work and in the discussion period I wrote down his claims against things I'd already published Point by point and he he ignored it a big promotion in by the government of Finland even after he did that and I think what an I am you know he wasn't worthy of it because of his ethics”
Competitive learning shows that humans and various mammals learn recognition categories (cell populations that fire selectively to spatial patterns of activation across feature-selective neurons), but experiences catastrophic forgetting when input patterns are numerous, statistically drifting, or overlapping—solving this requires the addition of top-down learned expectations to prevent catastrophic forgetting.
“competitive learning was good as far as it went and it responds well to sparse input environments well what does that mean in fact even deep learning doesn't experience catastrophic sparse input environment what it means that the input patterns aren't too close together in pattern space if you think of them as points in a multi-dimensional space and so I published a stable sparse learning theorem about competitive learning in 1976 showing if the input Vector space is sparse everything is good you learn Ice categories that are stable but I then realized that if there are too many input patterns for the uh way of instructed the network or the inputs that are presented through time are related to each other like you might have a statistical drift in changing inputs through time then previously learned categories got washed away by new learning so competitive learning experiences catastrophic forgetting”
The existence of the ART matching rule realized by top-down modulatory on-center-off-surround networks has been supported by psychological, anatomical, and physiological data in multiple species including bats, monkeys, and humans—indicating this circuit has been conserved through evolution.
“the existence of the prediction at an art matching rule that is realized by this kind of top-down modulatory on Center of surround network has been supported by psychological anatomical physiological data in multiple species including that Nobu osuga came to our department once and he was very excited about his new results about top-down matching and bats and the circuit obeyed the art matching rule that had been conserved at least from bats to humans and monkeys”
Grossberg wanted to continue his work and sought places with mathematical psychology programs; Stanford had such a program under William Estes (a great experimentalist who worked with B.F. Skinner), but Estes could not understand Grossberg's work on real-time autonomous adaptation in changing worlds, making Stanford not a good fit.
“I thought I can go there they'll get me but they didn't because bill was an experimentalist never had so what did he could do was to develop an urban model”
My magnum opus 'Conscious Mind, Resonant Brain: How Each Brain Makes a Mind' published in 2021 by Oxford University Press attempts to give a self-contained and non-technical overview and synthesis not only of my work and hundreds of collaborators over 50 years but also of hundreds of other scientists' work, winning the 2022 Prose Book Award in neuroscience.
“chapter 10 of my magnum opus which I haven't mentioned I should mention it because it it's called conscious mind resonant brain how each brain makes a mind it was published in 2021 by Oxford University president it tries to give a self-contained and non-technical overview and synthesis of not only my work and hundreds of collaborators over the past 50 some ideas but also explains and unifies the understanding of the work of hundreds of other scientists it's not just about me and I'm happy to say that the book won the 2022 Pro book award in neuroscience of the association of American Publishers”
The book has over 600 color figures, costs $17 for Kindle and $33 for hardcopy because Grossberg spent thousands of dollars of his own money to make it affordable, especially for young students; it is written to be self-contained and readable with chapters that can be read independently.
“it's Dirt Cheap even it's almost 800 pages with over 600 color figures the Kindle is just 17 bucks and the hard copy is just 33 bucks because I spent thousands of dollars my own money to make it affordable especially the young people like students”
At Rockefeller, Grossberg got his PhD developing proven global limit theorems about learning and memory; he received glowing reviews from both Mark Kac and John Paul Rota (both professors at Rockefeller), providing strong support for his future career; this shows how having advocates at critical moments can be decisive.
“I got my PhD from Rockefeller and then Mark Cox and John paularova who were both professors there wrote me glowing um reviews”
The real world is a blooming buzzing confusion where inputs are bombarded and nothing is sparse; everything is overlapping, which makes it impossible for sparse-input learning theories like competitive learning to solve realistic learning problems.
“we live in a blooming buzzing confusion where we're bombarded with inputs and nothing is sparse everything is overlapping”
Six different kinds of adaptive resonances support different aspects of consciousness and brain function: cognitive-emotional resonances (conscious feeling and knowing emotions), surface-shroud resonances (conscious seeing of visual objects), feature-category resonances (conscious recognition of learned visual objects), stream-shroud resonances (conscious hearing of auditory streams), spectral-pitch-timbre resonances (conscious recognition of auditory objects like music), and item-list resonances (conscious recognition of speech and language).
“I like distinguishing six different kind of adaptive resonances that I've studied um uh they support different aspects of Consciousness you were saying that's what's triggering this thought with different functions in different parts of our brains so I'll start with cognitive emotional resonance it's so not surprisingly they support conscious feeling of emotions and also knowing their Source oh I love that person you know the feeling and the target of the feeling but then there are five others”
My existential struggle was not 'what do I want to do' but 'how do I survive when I see things differently than what others say'—I knew what I saw in my discoveries before I had equations for them, and I had to trust my imagination and follow my heart while maintaining intellectual integrity.
“I think I might have been my first term a lot of people have existential struggles my struggle wasn't what I wanted to do which I was passionately committed to is how to Survivor yeah it was just like if someone says that's green but I know it's red I have a choice either I accept that it's red or I go crazy I knew what I saw I saw my discoveries before I had any equation for them and I followed my heart and my my imagination”
Honoré Matisse knew that all boundaries are invisible and many of the paintings Matisse published in 1905 are based on the observance that brains complete invisible boundaries to recognize objects—such boundary completion is a serious survival ability enabling detection of predators in partially camouflaged environments.
“I like to say that honoree Matisse whose work I love knew that all boundaries are invisible indeed there are many persons that we recognize by completing invisible boundaries haven't explained why boundaries are invisible in the boundary stream visibility Consciousness ability the property of surfaces which we have discussed a little and you know an example of completion of invisible boundaries in order to understand an object do you know the famous examples of a Dalmatian in snow”
Mind and self emerge from solving the stability-plasticity dilemma: the mind is the collective set of increasingly complex interacting awarenesses, abilities, and thoughts accumulated over a lifetime of learning, while the self is the evolving identity formed by attaching all this accumulated understanding to one's own person—both change continuously throughout life.
“and mind and self to come back to that are possible because ourselves the stability plasticity dilemma because as a result of that you can have a lifetime of learning that wall gets embedded in increasingly complex interacting awarenesses and uh abilities and thoughts about how the world works and collectively all of them together are our mind and as you go through life and you accumulate all of this understanding and attach it to your own person you have an evolving self so the self I have today is very different from the stuff I had as a boy or even when I was in my 30s or 40s I'm a quite a different being in some ways”
I have discussed how 14 different artists including Cézanne, Monet, and Rembrandt make paintings and how humans consciously see them, revealing the instinctive discoveries artists made about how minds consciously see that enable them to develop personal styles emphasizing some processes rather than others.
“I did stuff on how visual artists make paintings and how you'll must consciously see them so you know I'm really fascinated by how artists make paintings and how humans consciously see paintings and I discussed the work that I think it was 14 different artists including that case in Monet and Rembrandt”
If VLSI chips specialized to embody different intelligent capabilities could be created based on cortical principles, and combined in self-consistent ways, we could move toward general-purpose autonomous adaptive intelligence, particularly in algorithms and robots.
“if it could be embodied in a blsi ship certainly specialize to embody these different uh intelligent capabilities we can move toward general purpose autonomous adaptive intelligence and because they're chips they can go in everything from algorithms and machines to mobile robots because they're light it'll take a big big effort”
Detlef Bronk, president of Rockefeller Institute (who founded the Bronx as a borough), offered Grossberg a position based on the simple question 'If we funded you, would you want to come to Rockefeller?'—exemplifying how the right mentors at critical moments can shape scientific careers through trust rather than evaluation.
“the president with deadlift wrong p-r-o-n-k who uh was family founded the Bronx the borough of the Bronx in New York and colonial times he had previously been present in Johns Hopkins and I go into his fabulously elegantly furnished office just as the Sun is setting casting dramatic Shadows over the room and he sits me on this lovely sofa and he looks me in the eye and he said Steve if we funded you would you want to come to Rockefeller”
I was valedictorian at Stuyvesant High School in New York, a highly selective school, and at a graduation party heard from another student that he hated me because he wanted my ranking to go up by one—this competition for hundredths of decimal points drove home the need to stop learning for competition and start learning for love.
“coming out of New York where your ranking like it's Stuyvesant where I was valedictorian I don't know if you've heard of Stuyvesant but being some people think being felt Victorian status is the hardest thing you'll ever do but I was at a party after graduation and a guy was a little drunk came up to me it was for you know it was five percent Bronx science students to say the new garage parked and I said yeah I've been mad and he said no but I hated you my old years time it started soon I said why'd you pay me so well I wanted you to die because my grade my ranking would go up by one”
At Dartmouth I was offered a senior fellowship that allowed me to do research for a year instead of taking courses, which I needed to figure out my life direction—combining this with taking introduction to psychology, I had intellectual breakthroughs that set my entire life's work in motion.
“I really wanted to get the hell out of New York and so I went to Dartmouth because first they gave me the biggest scholarship but second they had something they call the senior Fellowship which meant in your senior year uh you could not take courses you could just do research now I didn't know what research was but I did know that I really really needed a year before I got out of college to figure out what to do with the rest of my life so when I went to Dartmouth as I indicated at the beginning I took introductory psychology and my head exploded and I started making discoveries”
Al Hastov, chairman of psychology, was a second mentor who later became dean and provost at Stanford; he nurtured my intellectual development and exemplified the balance between mathematics and psychology that enabled my subsequent work.
“and my other Mentor with Al hastov with chairman of psychology so I changed the math chairman of psychology went on to become a dean and Provost at Stanford he was a wonderful man so those are the two pillars of my work mathematics and psychology and I became the first joint major in mathematics and psychology at Darkness”
Mark Kac, a famous probability theorist and functional analyst who was Hungarian, was an advisor when I went to Rockefeller; during our interview lunch in a paneled room with art, his advice was 'Steve never grow old'—a message emphasizing staying young in spirit and intellectual energy.
“I had lunch with Mark Cox KAC a very famous probability theorist and functional analyst a fellow Hungarian and we had it in this amazing paneled room overlooking Gardens with art and uh paintings on the walls linen silverware anyway the big big bark told me was Steve never grow old that was our interview”
The speaker's primary gift is being able to see to the heart of data and understand what it means, combined with sufficient mathematical technique to carry out intuitive ideas in quantitative testable ways—this is why he obtained a PhD in mathematics and one professorship in mathematics and statistics.
“I think my my main gift is being able to see to the heart of data what what it means and then I have enough technique to carry out the program and that's one reason if not the main reason that I got a PhD in mathematics and one of my professorship professor mathematics and statistics so I had the tools to uh carry out my intuitive ideas in a quantitative and uh testable way”
I live in the data and follow anomalies obsessively: when seeing a fact I don't understand that interests me, I become 'consumed' until I can explain it in context; challenges to my models come from data itself—if presented with first-rate empirical contradictions, I would engage, but with small data samples one can explain any fact in 20 different ways, whereas explaining hundreds of facts in one way is the real test.
“I live in the data if someone would show me first I often would I'm always reading data at this stage in my life I'll often read data and I'll say oh that's just this or that's just that and I could see you know well to to write a paper on it it's a lot of effort but I know what the data means if I see a fact which is very unusual now that I don't know what the hell is going on I am consumed it will own me until I can put it in a context where I can explain it”
Thought experiments must be grounded in deep understanding of the target phenomena, not constructed in isolation—you cannot wake up and invent a thought experiment; the power comes from having previously grasped deep principles through study, then stepping back to see what simple ideas carry conjoint implications that generate insights.
“but you don't just wake up in the morning to say hey what's my latest thought experiment it's only after you already have a deep understanding of the targeted class of phenomena that you can step back and see well what are the simple ideas that if I carry them out as a logical story their conjoint implications you get to something you never thought you'd understand the first place”
I always knew I had a gift for imagining new worlds and working in complex domains, but realized I couldn't achieve world-class performance in all domains (music, visual art) without early specialization—I chose intellectual work where my 'gift for imagining other worlds' could sustain a lifetime of creative discovery.
“I always knew I had a gift um in fact when I was in public school uh my friend said I was born with a Golden Spoon in my mouth because I was good at everything it wasn't to be showy in fact and the second part of the sentence is and you're a nice guy so it wasn't something that I I felt was a reason to show off in fact I wish it just was good at everything I did”
A young mathematician at a catastrophe theory conference told me directly 'I don't care about data'—he wanted to apply catastrophe theory to biology without learning biology, trying to get something for nothing: good mathematical frameworks should be constrained to domains where their assumptions actually apply, and unfounded applications constitute intellectual failure.
“a young mathematician and he literally said to me isn't it wonderful that by knowing the elementary catastrophes I can write about biology but I don't have to learn any biologists it was really a case he said it he said very intelligent young man in case of trying to get something for nothing”
I achieved a functional understanding of consciousness that I never intended to seek—I was forced into it by being honest about my models and by the models' explanatory power, discovering that consciousness and related processes emerge from adaptive resonances that support learning, attention, and prediction.
“with this Foundation I was eventually forced into a functional understanding of Consciousness and I have to admit immediately I never try to understand Consciousness yes you don't wake up in the morning and say oh hey today I'm going to understand Consciousness you know never could happen just as I never thought when I started setting psychology and learning that I wouldn't be thrown into the neurodynamics behind learning you know these are the things that happen if you're honest and lucky enough to find true principles because they will never uh um disappoint you I call it the gift that keeps on giving”
Some people at Rockefeller wanted to throw me out, but failed, mostly because they didn't understand my work and felt unqualified—I survived because of support from Mark Kac and John Paularova who wrote glowing reviews, and then at Stanford Richard Atkinson (who became president of the National Academy of Sciences) and others champions my work.
“by people wanted to throw me out of Rockefeller um foreign reasons they failed but mostly because they didn't understand what I was doing and it made them feel uh you know unqualified but I got my PhD from Rockefeller and then Mark Cox and John paularova who were both professors there wrote me glowing um reviews”
My multidisciplinary approach across mathematics, psychology, neuroscience, and other fields emerged because the universality of subject matter forced me sometimes kicking and screaming into areas I never planned to study, but I've been true to the data and believe the majesty of evolution provides richness we must do our best to understand.
“it's because of the universality of the subject matter that I've Illustrated I've been LED sometimes Kicking and Screaming into areas I never thought I'd know anything about including consciousness in a way I've been forced into these fields by the power of the theory and and I feel very lucky to have been put in that position but you know I do think that I've been true to the data”
There have been vested interests opposed to the speaker's work—people who either lacked the preparation to compete with ART or were so invested in their current approaches that they wished the concept would go away rather than become the best game in town.
“but there were a lot of vested interests too would wish I'd go away because either they didn't have the preparation to compete although I don't think it in those terms or they were so in love with what they were currently doing that they uh wanted any concept that there wasn't this wasn't the best game in town to go bye-bye”
My modeling method cycles through analyzing scores or even hundreds of psychological experiments to discover design principles, deriving minimal models that embody those principles using Occam's razor and principles of parsimony, then showing the models explain more psychological data than the original hypotheses—creating self-consistent models that include previous models in more sophisticated form.
“the method showed how to discover design principles that underlie these large psychological databases and you might say well how does that happen and here I just have to quote the greatest of them all Albert Einstein it was a speculative leap if you live in hundreds of experiments long enough you can begin to see what the underlying principles and mechanisms are that generated those behaviors as emergent properties it's a gift it's a leap there is no algorithm for it”
The speaker distinguishes between consciousness understood as a verb (something constantly happening in real time) rather than as a noun (a single thing), which helps clarify how consciousness is related to the ongoing adaptive processes of the brain.
“it's very clear and it's easy to see how this sort of makes sense when you think of Consciousness as a verb rather than as a noun or or something that's constantly happening rather than just one single thing”
Attentional and orienting systems are computationally complementary like yin and yang, fitting together as puzzle pieces; they emerge in early development as a global symmetry breaking in morphogenesis, though this process is not yet fully understood and represents a major unsolved problem.
“so many of the processes I've understood a computationally complementary like yin yang and I think of them as emerging in early development as a kind of Global symmetry breaking in morphogenesis which I don't fully understand”
Young researchers too often only read the last 5-10 years of literature rather than foundational work published 30-50 years ago; rediscovering and reinventing existing results wastes time and talent, and when Grossberg notices someone rediscovering his work, he writes friendly letters pointing them to earlier results—good people acknowledge and engage, poor character practitioners ignore such corrections.
“I noticed just as often young investigators who know nothing about the work working on similar problems and I think the problem is that all too often today um investigators might think they only need to read the last five years of research or maybe maybe if the real scholar was 10 years but I made some of these discoveries 30 40 50 years ago and they're still holding yeah and um and I think that's okay no one can know everything but you know if I see someone's making a claim that is particularly irksome because I worked so hard to explain it much better than they were even doing now I'll write them applied and friendly letter saying hey you might be interested in knowing this result”
Last year I wrote a nautical article on music learning and conscious performance of lyrics and melodies with variable rhythms and beats, showing how musical competence co-opts previously existing foundational capabilities—all open access on my website.
“last year I wrote nautical Alec very much toward understanding the brain dynamics of Music Learning and conscious performance of lyrics and Melodies with variable rhythms and beats”
A new article 'How Children Learn to Understand Language Meanings: A Neural Model of Adult-Child Multimodal Interactions in Real Time' synthesizes many earlier discoveries—consciousness, learning, attention, perception, cognition, emotion—into unified understanding of how language meaning emerges from interaction, contrasting with ChatGPT which is just a lookup table with probabilistic prediction of next words, having no concept of meaning.
“just a few days ago to an article that I really liked a lot and it represents the synthesis of many of my earlier discoveries you know everything is integrated and the title is how children learn to understand language meanings colon and neural model of adult child multimodal interactions in real time”
Over 500 archival papers are available free on Grossberg's website (sites.bu.edu/SG) including the 1980 psychological review paper introducing the thought experiment explaining how systems autonomously learn to correct predictive errors, plus videos of keynote lectures and interviews, all open access.
“if people don't want to buy the book if you look at my 1980 paper in psychological review I have that and over 500 papers on my web page which is sites that's s-i-t-e-s Dot b u to Boston university.edu for Education slash Steve g s t e d e g if you go to sites.bu.edu slash DG that's my web page there are hundreds of papers 1980 you can download my psychological review paper for free”
Grossberg is currently writing another book for a general audience; he doesn't discuss unfinished projects because he might not complete them or might change direction if a new big idea emerges; he doesn't take on graduate students or postdocs at age 83 as it's unfair to commit them to 3-year projects when he could die.
“I'm writing another book for a general audience any is there anything you can say about the book or is that still just spoiler free no I I don't like talking about things I haven't yet done because I might not do them you know yeah I might change my mind”
In introductory economics at Dartmouth, Grossberg's final exam was so exceptional that his professor said it was at the level of a PhD thesis; his brother was at a different college and heard about this achievement, which made it harder for him to be the sibling of such an exceptional student.
“in my introductory economics course my final was so Sensational apparently that my teacher said it was at the level of a PhD thesis and my brother who was a different College heard about it which didn't make it easy for him to have me as a brother”
Grossberg knew he had a gift from childhood; his friend said he was born with a golden spoon in his mouth because he was good at everything and was a nice person; he was good at art, competitive award to study at Museum of Modern Art, learned piano quickly with great facility but realized classical music required starting younger and perfect hands he didn't have.
“I always knew I had a gift um in fact when I was in public school uh my friend said I was born with a Golden Spoon in my mouth because I was good at everything”
Grossberg has approximately 83,000 citations; if crediting all those who used his work, he would have 3-4 times more citations; citations are not the reason to work; staying productive requires studying problems rich enough with enough evidence to carry you to new and exciting ideas.
“you were saying I think was around 83 000 citations but if you look at the people who have been credited with a lot of the ideas I had first and developed and you estimate how many citations they got from using my work I would have three or four times that many citations but citations aren't the reason to work”
The speaker is called the 'Einstein of the mind' because he uses thought experiments to derive fundamental theories, just as Einstein derived special and general relativity from thought experiments; ART and his cognitive-emotional model were also derived from thought experiments.
“I'm called the Einstein of the mind because I use thought experiments just like Einstein derived both special relativity and general relativity from thought experiments like derived art from the thought experiment and I also derived my cognitive emotional model from a thought experiment and competitive networks from thought experiment”
The speaker has worked harder than one should ever work for knowledge—driven by the transition from competitive to intrinsic motivation, he pursued ideas with full commitment.
“so I I worked harder than you should ever work for knowledge but and then my other Mentor with Al hastov with chairman of psychology so I changed the math chairman of psychology went on to become a dean and Provost at Stanford he was a wonderful man so those are the two pillars of my work mathematics and psychology and I became the first joint major in mathematics and psychology at Darkness and in my senior Fellowship here I worked like a dog to further develop the discoveries I started making as a freshman”
I am currently 83 years old; my mother lived to 101 and I have serious asthma but no family history of dementia, so if the asthma doesn't kill me I expect to be productive for maybe 10 more years; I am retired since 2019 and no longer seek research grants, allowing me to work solo on whatever ideas capture my attention.
“I I am 83. my mom lived to 101. Jesus I have serious asthma so if miasma doesn't kill me there's no history of Dementia in the family if if I don't get sick I'll probably be productive for maybe 10 more years who knows um what I'm doing now is writing another book for a general audience”
Rockefeller's president asked the speaker directly: 'If we funded you, would you want to come to Rockefeller?' in a dramatically lit office overlooking gardens, representing an unusual and direct approach to recruitment.
“I go into his fabulously elegantly furnished office just as the Sun is setting casting dramatic Shadows over the room and he sits me on this lovely sofa and he looks me in the eye and he said Steve if we funded you would you want to come to Rockefeller that was my interview and I said yes”