Presenting our work today at #CogSci2026 🎉 Our findings reveal interesting similarities, and important differences, in how humans and modern AI systems explore causal structures. If you're at the conference, come by and say hello! Today 5:30–7:00pm Reasoning session, P3-L-104

Mandana Samiei@mandanas.bsky.social · 2mo ago

We recently investigated how humans and AI reason through causal rules: arxiv.org/abs/2606.06464 Accepted at #CogSci2026. We gave adults and LLMs agency to actively experiment using a Blicket causal game, breaking down their process into the iterative loop shown in the following figure:

1/2) Please let's all remember that GWT is a theoy of access consciousness (what is available for verbal reasoning) and not phenomenal consciousness (subjective experience): philpapers.org/rec/BLOOAC Obviously, in this study access consciousness is literally what they test in the LLMs. #NeuroAI 🧪

Anthropic {bot}@anthropicbot.bsky.social · last mo.

New Anthropic research: A global workspace in language models. Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with. (1/2)

Check out our new work: how do humans and LLMs explore causal structures differently?

Mandana Samiei@mandanas.bsky.social · 2mo ago

We recently investigated how humans and AI reason through causal rules: arxiv.org/abs/2606.06464 Accepted at #CogSci2026. We gave adults and LLMs agency to actively experiment using a Blicket causal game, breaking down their process into the iterative loop shown in the following figure:

We recently investigated how humans and AI reason through causal rules: arxiv.org/abs/2606.06464 Accepted at #CogSci2026. We gave adults and LLMs agency to actively experiment using a Blicket causal game, breaking down their process into the iterative loop shown in the following figure:

Two new preprints from the lab, First up, causal exploration in humans and LLM's. arxiv.org Human Adults and LLMs as Scientists: Who Benefits from Active Exploration?

Very much looking forward to this!

Mark Ho@markkho.bsky.social · 2mo ago

How do human minds make sense of big, messy problems? 😵‍💫🌀 How do we distill complexity into something simple enough to solve? 🤔💡 We’ll be tackling these questions (and more!) at two workshops on task representations, abstractions, and construals #CogSci2026 #CCN2026 🧵 framing-the-problem.github.io

How do brains plan actions towards goals? To get at this question we studied mice navigating complex mazes as goals changed on every trial 🧵 Work with @thomasakam.bsky.social @behrenstimb.bsky.social @kristorpjensen.bsky.social now on BioRxiv: www.biorxiv.org/content/10.6...

Perspective piece from Blake @tyrellturing.bsky.social @mandanas.bsky.social 👇

Blake Richards@tyrellturing.bsky.social · 2mo ago

New perspective piece with @mandanas.bsky.social: We argue, based on LLMs and old connectionist theories, that schemas shouldn't be viewed as distinct from semantic or episodic memories. They're just one end of a detailed-to-abstract memory spectrum: www.cell.com/neuron/fullt... #NeuroAI 🧠📈 🧪

New perspective piece with @mandanas.bsky.social: We argue, based on LLMs and old connectionist theories, that schemas shouldn't be viewed as distinct from semantic or episodic memories. They're just one end of a detailed-to-abstract memory spectrum: www.cell.com/neuron/fullt... #NeuroAI 🧠📈 🧪

The schema spectrum: Emergent structures and levels of abstraction in AI and the brain

Motivated by generative AI, Samiei et al. argue against classical models that treat schemas as distinct memory structures. Instead, they propose that schemas are merely a conceptual tool describing ho...

cell.com

Excited to see our perspective out today! We argue that schemas may not be fundamentally distinct from semantic and episodic memories, but instead lie along a spectrum of abstraction, from detailed episodic experiences to highly abstract knowledge. See more: doi.org/10.1016/j.ne...

How can machines achieve the kind of flexible learning mastered by our brains? That’s what our own @neurokim.bsky.social is trying to find out! Learn more about ARNI’s work on natural and artificial intelligence: youtu.be/gM8YCaiAWrc?... #BrainAwarenessWeek

ARNI: Memory Driven Learning with Kim Stachenfeld

YouTube video by Columbia Engineering

youtu.be

Check out the Call for Workshops at RLC this year. There is still nearly a month before the deadline, we are looking forward to your proposals!

Taylor W. Killian@twkillian.bsky.social · 6mo ago

We're thrilled to share that the Call for Workshops for this year's @rl-conference.bsky.social is now live! As Workshop co-chair (alongside the wonderful Raksha Kumaraswamy and @claireve.bsky.social) we are looking forward to seeing the proposals for workshops that we receive. LINK IN NEXT POST

What is the relationship between memorization and generalization in AI? Is there a fundamental tradeoff? In infinitefaculty.substack.com/p/memorizati... I’ve reviewed some of the evolving perspectives on memorization & generalization in machine learning, from classic perspectives through LLMs.

Memorization vs. generalization in deep learning: implicit biases, benign overfitting, and more

Or: how I learned to stop worrying and love the memorization

infinitefaculty.substack.com

Excited to share that our work on Visual Symbolic Mechanisms has been accepted to ICLR! 🌴🇧🇷

Declan Campbell@thisisadax.bsky.social · 6mo ago

The visual world is composed of objects, and those objects are composed of features. But do VLMs exploit this compositional structure when processing multi-object scenes? In our 🆒🆕 #ICLR2026 paper, we find they do – via emergent symbolic mechanisms for visual binding. 🧵👇

Huge congratulations! Already looking forward to the breakthroughs you'll lead!

Dan Levenstein@dlevenstein.bsky.social · last yr.

Thrilled to announce I'll be starting my own neuro-theory lab, as an Assistant Professor at @yaleneuro.bsky.social @wutsaiyale.bsky.social this Fall! My group will study offline learning in the sleeping brain: how neural activity self-organizes during sleep and the computations it performs. 🧵

Great deep dive into hierarchical reinforcement learning, essential reading for anyone exploring scalable, structured agents. Shoutout to the authors!

Martin Klissarov@martinklissarov.bsky.social · last yr.

As AI agents face increasingly long and complex tasks, decomposing them into subtasks becomes increasingly appealing. But how do we discover such temporal structure? Hierarchical RL provides a natural formalism-yet many questions remain open. Here's our overview of the field🧵

Many LMs default to disjunctive inferences—even when given conjunctive evidence. Unlike children, LMs’ exploration is shaped by the underlying causal rule. Could promoting child-like curiosity help LMs reason more effectively about causality? More details: arxiv.org/abs/2505.09614

Mandana Samiei@mandanas.bsky.social · last yr.

A great collab with former labmates @agx-chen.bsky.social & @dongyanl1n.bsky.social. Interesting limitation in LMs: strong disjunctive bias leads to poor performance on conjunctive causal inference tasks. Mirrors adult human biases--possibly a byproduct of training data prior.

A great collab with former labmates @agx-chen.bsky.social & @dongyanl1n.bsky.social. Interesting limitation in LMs: strong disjunctive bias leads to poor performance on conjunctive causal inference tasks. Mirrors adult human biases--possibly a byproduct of training data prior.

Anthony GX-Chen@agx-chen.bsky.social · last yr.

Language model (LM) agents are all the rage now—but they may exhibit cognitive biases when inferring causal relationships! We evaluate LMs on a cognitive task to find: - LMs struggle with certain simple causal relationships - They show biases similar to human adults (but not children) 🧵⬇️

Just over a week since I defended my 🤖+🧠PhD thesis, and the feeling is just sinking in. Extremely grateful to @tyrellturing.bsky.social for supporting me through this amazing journey! 🙏 Big thanks to all members of the LiNC lab, and colleagues at mcgill University and @mila-quebec.bsky.social. ❤️😁

The slides of my NeurIPS lecture "From Diffusion Models to Schrödinger Bridges - Generative Modeling meets Optimal Transport" can be found here drive.google.com/file/d/1eLa3...

I gave a talk on Compositional World Models at NeurIPS last week 🌐 The recording is now online: neurips.cc/virtual/2024... (for registered attendees; starts at 6:06:00) Workshop: compositional-learning.github.io

🔊 Super excited to announce the first ever Frontiers of Probabilistic Inference: Learning meets Sampling workshop at #ICLR2025 @iclr-conf.bsky.social! 🔗 website: sites.google.com/view/fpiwork... 🔥 Call for papers: sites.google.com/view/fpiwork... more details in thread below👇 🧵

If you review for a #ML conference like @iclr-conf.bsky.social or @neuripsconf.bsky.social, YOU HAVE A RESPONSIBILITY TO REPLY TO THE AUTHORS. If the rebuttal doesn't address your concerns explain why. But giving a score of 2-3 then ghosting the authors is super rude. I say this as an AC. #MLSky

Now that @jeffclune.bsky.social and @joelbot3000.bsky.social are here, time for an Open-Endedness starter pack. go.bsky.app/MdVxrtD

4/ I think the proper definition is that #NeuroAI is the realization of the original promise of cybernetics and connectionism! en.wikipedia.org/wiki/Cyberne... en.wikipedia.org/wiki/Connect... It is a general science of intelligence focussed on parallel distributed systems, control, and learning.

1/ I work in #NeuroAI, a growing field of research, which many people have only the haziest conception of... As way of introduction to this research approach, I'll provide here a very short thread outlining the definition of the field I gave recently at our BRAIN NeuroAI workshop at the NIH. 🧠📈

I wish that I was able to convince the students in my class that the ultimate goal of science isn't accumulation of facts but the compression of facts into theories. Mel summarizes it well here: www.nature.com/articles/nn1... Unfortunately, I think this is a minority viewpoint within neuroscience.