Check out my free video series about what's missing in AI and Neuroscience

Support the show to get full episodes and join the Discord community.

Jeffrey Bowers is a psychologist and professor at the University of Bristol. As you know, many of my previous guests are in the business of comparing brain activity to the activity of units in artificial neural network models, when humans or animals and the models are performing the same tasks. And a big story that has emerged over the past decade or so is that there's a remarkable similarity between the activities and representations in brains and models. This was originally found in object categorization tasks, where the goal is to name the object shown in a given image, where researchers have compared the activity in the models good at doing that to the activity in the parts of our brains good at doing that. It's been found in various other tasks using various other models and analyses, many of which we've discussed on previous episodes, and more recently a similar story has emerged regarding a similarity between language-related activity in our brains and the activity in large language models. Namely, the ability of our brains to predict an upcoming word can been correlated with the models ability to predict an upcoming word. So the word is that these deep learning type models are the best models of how our brains and cognition work.

However, this is where Jeff Bowers comes in and raises the psychology flag, so to speak. His message is that these predictive approaches to comparing artificial and biological cognition aren't enough, and can mask important differences between them. And what we need to do is start performing more hypothesis driven tests like those performed in psychology, for example, to ask whether the models are indeed solving tasks like our brains and minds do. Jeff and his group, among others, have been doing just that are discovering differences in models and minds that may be important if we want to use models to understand minds. We discuss some of his work and thoughts in this regard, and a lot more.

• Website
• Twitter: @jeffrey_bowers
• Related papers:
  • Deep Problems with Neural Network Models of Human Vision.
  • eu-w2.wpmucdn.com/blogs.bristol.ac.uk/dist/b/403/files/2017/11/bowers-tics-2017.pdf">Parallel Distributed Processing Theory in the Age of Deep Networks.
  • Successes and critical failures of neural networks in capturing human-like speech recognition.

0:00 - Intro 3:52 - Testing neural networks 5:35 - Neuro-AI needs psychology 23:36 - Experiments in AI and neuroscience 23:51 - Why build networks like our minds? 44:55 - Vision problem spaces, solution spaces, training data 55:45 - Do we implement algorithms? 1:01:33 - Relational and combinatorial cognition 1:06:17 - Comparing representations in different networks 1:12:31 - Large language models 1:21:10 - Teaching LLMs nonsense languages

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Gary Lupyan runs the Lupyan Lab at University of Wisconsin, Madison, where he studies how language and cognition are related. In some ways, this is a continuation of the conversation I had last episode with Ellie Pavlick, in that we  partly continue to discuss large language models. But Gary is more focused on how language, and naming things, categorizing things, changes our cognition related those things. How does naming something change our perception of it, and so on. He's interested in how concepts come about, how they map onto language. So we talk about some of his work and ideas related to those topics.

And we actually start the discussion with some of Gary's work related the variability of individual humans' phenomenal experience, and how that affects our individual cognition. For instance, some people are more visual thinkers, others are more verbal, and there seems to be an appreciable spectrum of differences that Gary is beginning to experimentally test.

• Lupyan Lab.
• Twitter: @glupyan.
• Related papers:
  • Hidden Differences in Phenomenal Experience.
  • Verbal interference paradigms: A systematic review investigating the role of language in cognition.
• Gary mentioned Richard Feynman's Ways of Thinking video.
• Gary and Andy Clark's Aeon article: Super-cooperators.

0:00 - Intro 2:36 - Words and communication 14:10 - Phenomenal variability 26:24 - Co-operating minds 38:11 - Large language models 40:40 - Neuro-symbolic AI, scale 44:43 - How LLMs have changed Gary's thoughts about language 49:26 - Meaning, grounding, and language 54:26 - Development of language 58:53 - Symbols and emergence 1:03:20 - Language evolution in the LLM era 1:08:05 - Concepts 1:11:17 - How special is language? 1:18:08 - AGI

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Ellie Pavlick runs her Language Understanding and Representation Lab at Brown University, where she studies lots of topics related to language. In AI, large language models, sometimes called foundation models, are all the rage these days, with their ability to generate convincing language, although they still make plenty of mistakes. One of the things Ellie is interested in is how these models work, what kinds of representations are being generated in them to produce the language they produce. So we discuss how she's going about studying these models. For example, probing them to see whether something symbolic-like might be implemented in the models, even though they are the deep learning neural network type, which aren't suppose to be able to work in a symbol-like manner. We also discuss whether grounding is required for language understanding - that is, whether a model that produces language well needs to connect with the real world to actually understand the text it generates. We talk about what language is for, the current limitations of large language models, how the models compare to humans, and a lot more.

• Language Understanding and Representation Lab
• Twitter: @Brown_NLP
• Related papers
  • Semantic Structure in Deep Learning.
  • 1.23.pdf">Pretraining on Interactions for Learning Grounded Affordance Representations.
  • Mapping Language Models to Grounded Conceptual Spaces.

0:00 - Intro 2:34 - Will LLMs make us dumb? 9:01 - Evolution of language 17:10 - Changing views on language 22:39 - Semantics, grounding, meaning 37:40 - LLMs, humans, and prediction 41:19 - How to evaluate LLMs 51:08 - Structure, semantics, and symbols in models 1:00:08 - Dimensionality 1:02:08 - Limitations of LLMs 1:07:47 - What do linguists think? 1:14:23 - What is language for?

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Earl Miller runs the Miller Lab at MIT, where he studies how our brains carry out our executive functions, like working memory, attention, and decision-making. In particular he is interested in the role of the prefrontal cortex and how it coordinates with other brain areas to carry out these functions. During this episode, we talk broadly about how neuroscience has changed during Earl's career, and how his own thoughts have changed. One thing we focus on is the increasing appreciation of brain oscillations for our cognition.

Recently on BI we've discussed oscillations quite a bit. In episode 153, Carolyn Dicey-Jennings discussed her philosophical ideas relating attention to the notion of the self, and she leans a lot on Earl's research to make that argument.  In episode 160, Ole Jensen discussed his work in humans showing that  low frequency oscillations exert a top-down control on incoming sensory stimuli, and this is directly in agreement with Earl's work over many years in nonhuman primates. So we continue that discussion relating low-frequency oscillations to executive control. We also discuss a new concept Earl has developed called spatial computing, which is an account of how brain oscillations can dictate where in various brain areas neural activity be on or off, and hence contribute or not to ongoing mental function. We also discuss working memory in particular, and a host of related topics.

• Miller lab.
• Twitter: @MillerLabMIT.
• Related papers:
  • Cohen-20011.pdf">An integrative theory of prefrontal cortex function. Annual Review of Neuroscience.
  • and-Miller-JOCN-2023.pdf">Working Memory Is Complex and Dynamic, Like Your Thoughts.
  • Waves-PLOS-Comp-Bio-2022.pdf">Traveling waves in the prefrontal cortex during working memory.

0:00 - Intro 6:22 - Evolution of Earl's thinking 14:58 - Role of the prefrontal cortex 25:21 - Spatial computing 32:51 - Homunculus problem 35:34 - Self 37:40 - Dimensionality and thought 46:13 - Reductionism 47:38 - Working memory and capacity 1:01:45 - Capacity as a principle 1:05:44 - Silent synapses 1:10:16 - Subspaces in dynamics

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Hugo Spiers runs the Spiers Lab at University College London. In general Hugo is interested in understanding spatial cognition, like navigation, in relation to other processes like planning and goal-related behavior, and how brain areas like the hippocampus and prefrontal cortex coordinate these cognitive functions. So, in this episode, we discuss a range of his research and thoughts around those topics. You may have heard about the studies he's been involved with for years, regarding London taxi drivers and how their hippocampus changes as a result of their grueling efforts to memorize how to best navigate London. We talk about that, we discuss the concept of a schema, which is roughly an abstracted form of knowledge that helps you know how to behave in different environments. Probably the most common example is that we all have a schema for eating at a restaurant, independent of which restaurant we visit, we know about servers, and menus, and so on. Hugo is interested in spatial schemas, for things like navigating a new city you haven't visited. Hugo describes his work using reinforcement learning methods to compare how humans and animals solve navigation tasks. And finally we talk about the video game Hugo has been using to collect vast amount of data related to navigation, to answer questions like how our navigation ability changes over our lifetimes, the different factors that seem to matter more for our navigation skills, and so on.

• Spiers Lab.
• Twitter: @hugospiers.
• Related papers
  • Predictive maps in rats and humans for spatial navigation.
  • From cognitive maps to spatial schemas.
  • London taxi drivers: A review of neurocognitive studies and an exploration of how they build their cognitive map of London.
  • Explaining World-Wide Variation in Navigation Ability from Millions of People: Citizen Science Project Sea Hero Quest.

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Ole Jensen is co-director of the Centre for Human Brain Health at University of Birmingham, where he runs his Neuronal Oscillations Group lab. Ole is interested in how the oscillations in our brains affect our cognition by helping to shape the spiking patterns of neurons, and by helping to allocate resources to parts of our brains that are relevant for whatever ongoing behaviors we're performing in different contexts. People have been studying oscillations for decades, finding that different frequencies of oscillations have been linked to a bunch of different cognitive functions. Some of what we discuss today is Ole's work on alpha oscillations, which are around 10 hertz, so 10 oscillations per second. The overarching story is that alpha oscillations are thought to inhibit or disrupt processing in brain areas that aren't needed during a given behavior. And therefore by disrupting everything that's not needed, resources are allocated to the brain areas that are needed. We discuss his work in the vein on attention - you may remember the episode with Carolyn Dicey-Jennings, and her ideas about how findings like Ole's are evidence we all have selves. We also talk about the role of alpha rhythms for working memory, for moving our eyes, and for previewing what we're about to look at before we move our eyes, and more broadly we discuss the role of oscillations in cognition in general, and of course what this might mean for developing better artificial intelligence.

• The Neuronal Oscillations Group.
• Twitter: @neuosc.
• Related papers
  • Shaping functional architecture by oscillatory alpha activity: gating by inhibition
  • FEF-Controlled Alpha Delay Activity Precedes Stimulus-Induced Gamma-Band Activity in Visual Cortex
  • The theta-gamma neural code
  • A pipelining mechanism supporting previewing during visual exploration and reading.
  • Specific lexico-semantic predictions are associated with unique spatial and temporal patterns of neural activity.

0:00 - Intro 2:58 - Oscillations import over the years 5:51 - Oscillations big picture 17:62 - Oscillations vs. traveling waves 22:00 - Oscillations and algorithms 28:53 - Alpha oscillations and working memory 44:46 - Alpha as the controller 48:55 - Frequency tagging 52:49 - Timing of attention 57:41 - Pipelining neural processing 1:03:38 - Previewing during reading 1:15:50 - Previewing, prediction, and large language models 1:24:27 - Dyslexia

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Chris Summerfield runs the Human Information Processing Lab at University of Oxford, and he's a research scientist at Deepmind. You may remember him from episode 95 with Sam Gershman, when we discussed ideas around the usefulness of neuroscience and psychology for AI. Since then, Chris has released his book, Natural General Intelligence: How understanding the brain can help us build AI. In the book, Chris makes the case that inspiration and communication between the cognitive sciences and AI is hindered by the different languages each field speaks. But in reality, there has always been and still is a lot of overlap and convergence about ideas of computation and intelligence, and he illustrates this using tons of historical and modern examples.

• Human Information Processing Lab.
• Twitter: @summerfieldlab.
• Book: Natural General Intelligence: How understanding the brain can help us build AI.
• Other books mentioned:
  • Are We Smart Enough to Know How Smart Animals Are? by Frans de Waal
  • The Mind is Flat by Nick Chater.

0:00 - Intro 2:20 - Natural General Intelligence 8:05 - AI and Neuro interaction 21:42 - How to build AI 25:54 - Umwelts and affordances 32:07 - Different kind of intelligence 39:16 - Ecological validity and AI 48:30 - Is reward enough? 1:05:14 - Beyond brains 1:15:10 - Large language models and brains

Check out my free video series about what's missing in AI and Neuroscience

Support the show to get full episodes and join the Discord community.

Paul Rosenbloom is Professor Emeritus of Computer Science at the University of Southern California. In the early 1980s, Paul , along with John Laird and the early AI pioneer Alan Newell, developed one the earliest and best know cognitive architectures called SOAR. A cognitive architecture, as Paul defines it, is a model of the fixed structures and processes underlying minds, and in Paul's case the human mind. And SOAR was aimed at generating general intelligence. He doesn't work on SOAR any more, although SOAR is still alive and well in the hands of his old partner John Laird. He did go on to develop another cognitive architecture, called Sigma, and in the intervening years between those projects, among other things Paul stepped back and explored how our various scientific domains are related, and how computing itself should be considered a great scientific domain. That's in his book On Computing: The Fourth Great Scientific Domain.

He also helped develop the Common Model of Cognition, which isn't a cognitive architecture itself, but instead a theoretical model meant to generate consensus regarding the minimal components for a human-like mind. The idea is roughly to create a shared language and framework among cognitive architecture researchers, so the field can , so that whatever cognitive architecture you work on, you have a basis to compare it to, and can communicate effectively among your peers.

All of what I just said, and much of what we discuss, can be found in Paul's memoir, 31-22.pdf?dl=0">In Search of Insight: My Life as an Architectural Explorer.

• Paul's website.
• Related papers
  • Working memoir: 31-22.pdf?dl=0">In Search of Insight: My Life as an Architectural Explorer.
  • Book: On Computing: The Fourth Great Scientific Domain.
  • A Standard Model of the Mind: Toward a Common Computational Framework across Artificial Intelligence, Cognitive Science, Neuroscience, and Robotics.
  • Analysis of the human connectome data supports the notion of a “Common Model of Cognition” for human and human-like intelligence across domains.
  • Common Model of Cognition Bulletin.

0:00 - Intro 3:26 - A career of exploration 7:00 - Alan Newell 14:47 - Relational model and dichotomic maps 24:22 - Cognitive architectures 28:31 - SOAR cognitive architecture 41:14 - Sigma cognitive architecture 43:58 - SOAR vs. Sigma 53:06 - Cognitive architecture community 55:31 - Common model of cognition 1:11:13 - What's missing from the common model 1:17:48 - Brains vs. cognitive architectures 1:21:22 - Mapping the common model onto the brain 1:24:50 - Deep learning 1:30:23 - AGI

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Sarah Robins is a philosopher at the University of Kansas, one a growing handful of philosophers specializing in memory. Much of her work focuses on memory traces, which is roughly the idea that somehow our memories leave a trace in our minds. We discuss memory traces themselves and how they relate to the engram (see BI 126 Randy Gallistel: Where Is the Engram?, and BI 127 Tomás Ryan: Memory, Instinct, and Forgetting).

Psychology has divided memories into many categories - the taxonomy of memory. Sarah and I discuss how memory traces may cross-cut those categories, suggesting we may need to re-think our current ontology and taxonomy of memory.

We discuss a couple challenges to the idea of a stable memory trace in the brain. Neural dynamics is the notion that all our molecules and synapses are constantly changing and being recycled. Memory consolidation refers to the process of transferring our memory traces from an early unstable version to a more stable long-term version in a different part of the brain. Sarah thinks neither challenge poses a real threat to the idea

We also discuss the impact of optogenetics on the philosophy and neuroscience and memory, the debate about whether memory and imagination are essentially the same thing, whether memory's function is future oriented, and whether we want to build AI with our often faulty human-like memory or with perfect memory.

• Sarah's website.
• Twitter: @SarahKRobins.
• Related papers:
  • Her Memory chapter, with Felipe de Brigard, in the book Mind, Cognition, and Neuroscience: A Philosophical Introduction.
  • 5fb2-4eed-b8b6-7899b2613e1c.filesusr.com/ugd/15e503_442ce387006d4c50b497b940a3eded1f.docx?dn=Memory%20and%20Optogenetic%20Intervention.docx">Memory and Optogenetic Intervention: Separating the engram from the ecphory.
  • Stable Engrams and Neural Dynamics.

0:00 - Intro 4:18 - Philosophy of memory 5:10 - Making a move 6:55 - State of philosophy of memory 11:19 - Memory traces or the engram 20:44 - Taxonomy of memory 25:50 - Cognitive ontologies, neuroscience, and psychology 29:39 - Optogenetics 33:48 - Memory traces vs. neural dynamics and consolidation 40:32 - What is the boundary of a memory? 43:00 - Process philosophy and memory 45:07 - Memory vs. imagination 49:40 - Constructivist view of memory and imagination 54:05 - Is memory for the future? 58:00 - Memory errors and intelligence 1:00:42 - Memory and AI 1:06:20 - Creativity and memory errors

Support the show to get full episodes and join the Discord community.

Check out my free video series about what's missing in AI and Neuroscience

Mariam Aly runs the Aly lab at Columbia University, where she studies the interaction of memory, attention, and perception in brain regions like the hippocampus. The short story is that memory affects our perceptions, attention affects our memories, memories affect our attention, and these effects have signatures in neural activity measurements in our hippocampus and other brain areas. We discuss her experiments testing the nature of those interactions. We also discuss a particularly difficult stretch in Mariam's graduate school years, and how she now prioritizes her mental health.

• Aly Lab.
• Twitter: @mariam_s_aly.
• Related papers
  • Attention promotes episodic encoding by stabilizing hippocampal representations.
  • The medial temporal lobe is critical for spatial relational perception.
  • Cholinergic modulation of hippocampally mediated attention and perception.
  • Preparation for upcoming attentional states in the hippocampus and medial prefrontal cortex.
  • How hippocampal memory shapes, and is shaped by, attention.
  • Attentional fluctuations and the temporal organization of memory.

0:00 - Intro 3:50 - Mariam's background 9:32 - Hippocampus history and current science 12:34 - hippocampus and perception 13:42 - Relational information 18:30 - How much memory is explicit? 22:32 - How attention affects hippocampus 32:40 - fMRI levels vs. stability 39:04 - How is hippocampus necessary for attention 57:00 - How much does attention affect memory? 1:02:24 - How memory affects attention 1:06:50 - Attention and memory relation big picture 1:07:42 - Current state of memory and attention 1:12:12 - Modularity 1:17:52 - Practical advice to improve attention/memory 1:21:22 - Mariam's challenges