Yang Wu

philosophymindscience.org/index.php/ph...

Enjoyed writing this with the wonderful philosopher Bill Ramsey

What does it mean for a system to compute? What does it mean for a system to compute?, Wolpert, David H, Korbel, Jan

What does it mean for a system to compute?
doi.org/10.1088/2632...

Huge congrats!! Excited to read it!
Been waiting for this since I heard you mentioned these experiments on podcasts

🚨🚨New Preprint Alert!🚨🚨

www.biorxiv.org/content/10.6...

Animal learning is painfully slow (at least initially). Yet, well trained animals can learn very fast, sometimes displaying few-shot inference. How does this transition occur?

RNA, the Epicenter of Genetic Information The origin story and emergence of molecular biology is muddled. The early triumphs in bacterial genetics and the complexity of animal and plant genomes complicate an intricate history. This book docum...

I just learned that Mattick & Amaral's book about RNA is availabe in an online version:
www.ncbi.nlm.nih.gov/books/NBK595...
Highly recommended! Not just for the science but also for the history of science aspect.

Apes Share Human Ability to Imagine YouTube video by Johns Hopkins University

Imagination in bonobos!

I am thrilled to share a new paper w/ Amalia Bastos, out now in @science.org

We provide the first experimental evidence that a nonhuman animal can follow along a pretend scenario & track imaginary objects. Work w/ Kanzi, the bonobo, at Ape Initiative

youtu.be/NUSHcQQz2Ko

"How much of the brain's learned algorithms depend on the fact it is a brain?" arxiv.org/abs/2601.02063 The brain is a neural network, but also a biological organ (unlike artificial neural networks). How much does this matter to cognition?

🧠 How do transformers learn relational reasoning? We trained small transformers on transitive inference (if A>B and B>C, then A>C) and discovered striking differences between learning paradigms. Our latest work reveals when and why AI systems generalize beyond training data 🤖

This Gershman and Ullman paper is just so cool in showing that humans really can't not view correlation without assuming causality. Summary of my quixotic life in neuroscience: gershmanlab.com/pubs/Gershma...

How do we achieve few-shot generalization? New work led by @fabianrenz.bsky.social dives into the role of replay in learning and using structure to generalize reward. Dream team effort with Shany Grossman @nathanieldaw.bsky.social Peter Dayan & @doellerlab.bsky.social
www.biorxiv.org/content/10.6...

A unifying account of replay as context-driven memory reactivation A context-driven memory model simulates a wide range of characteristics of waking and sleeping hippocampal replay, providing a new account of how and why replay occurs.

Really thrilled that this paper led by @neurozz.bsky.social is now published in its final version in @elife.bsky.social!!

This is a memory-focused (as opposed to RL-focused) account of the detailed characteristics of forward and backward awake and sleep replay!

elifesciences.org/articles/99931

Super cooool!

To sum:

A child waves a stick, yelling:“behold my wand!”. A teacher holds up an apple and says “this is the sun”. A parent points out two small carrots next to a big carrot and nudges his kids: “that’s us”.

Such acts are common, intuitive, and useful.

𝖡𝗋𝗂𝖾𝖿 𝗋𝖾𝗏𝗂𝖾𝗐 𝗐𝗂𝗍𝗁 𝖣𝖺𝗇 𝖡𝗎𝗌𝗁, 𝗈𝖿 𝗇𝖾𝗎𝗋𝖺𝗅 𝗌𝖾𝗊𝗎𝖾𝗇𝖼𝖾𝗌 𝗂𝗇 𝖿𝗅𝗒𝗂𝗇𝗀 𝖻𝖺𝗍𝗌 (𝐫𝐞𝐩𝐥𝐚𝐲 𝖺𝗇𝖽 𝐭𝐡𝐞𝐭𝐚 𝐬𝐰𝐞𝐞𝐩𝐬!) 𝗂𝗇 𝗍𝗐𝗈 𝗋𝖾𝖼𝖾𝗇𝗍 𝗉𝖺𝗉𝖾𝗋𝗌 𝖻𝗒 𝖠𝗇𝗀𝖾𝗅𝗈 𝖥𝗈𝗋𝗅𝗂, 𝖶𝗎𝖽𝗂 𝖥𝖺𝗇, 𝖪𝖾𝗏𝗂𝗇 𝖰𝗂 & 𝖬𝗂𝖼𝗁𝖺𝖾𝗅 𝖸𝖺𝗋𝗍𝗌𝖾𝗏, 𝖺𝗇𝖽 𝖻𝗒 𝖳𝖺𝗆𝗂𝗋 𝖤𝗅𝗂𝖺𝗏, 𝖲𝗁𝗂𝗋 𝖬𝖺𝗂𝗆𝗈𝗇, 𝖠𝗒𝖾𝗅𝖾𝗍 𝖲𝖺𝗋𝖾𝗅, 𝖲𝗁𝖺𝗄𝖾𝖽 𝖯𝖺𝗅𝗀𝗂, 𝖫𝗂𝗈𝗋𝖺 𝖫𝖺𝗌 & 𝖭𝖺𝖼𝗁𝗎𝗆 𝖴𝗅𝖺𝗇𝗈𝗏𝗌𝗄𝗒: authors.elsevier.com/a/1lpaB3QW8S...

Merry Christmas guys!

I ll recommend this paper. Read it once, like it a lot. Not fully grasp all details yet, I ll probably present it at Allen. Really like their previous Trial-Matching one arxiv.org/abs/2306.03603 . We are using/extending some of these OT ideas at Allen for fitting neural data. Inspiring work.

Ohh I see.. ty!

wait I never knew this lol

Excited to announce a new book telling the story of mathematical approaches to studying the mind, from the origins of cognitive science to modern AI! The Laws of Thought will be published in February and is available for pre-order now.

They certainly can be model systems, and model organisms are a type of model system. I think your points about serving as test cases, or as how-possible (or even how-actually) models, are exemplary of such a role as a model system.

I would add to this that we should be careful not to use psychological tests developed for measuring *human* qualities in *humans* and then conclude machines have human qualities. This is especially problematic with LLM and text based tests, as they are trained to simulate (human) text.

Subjective functions Where do objective functions come from? How do we select what goals to pursue? Human intelligence is adept at synthesizing new objective functions on the fly. How does this work, and can we endow arti...

Goal selection through the lens of subjective functions:
arxiv.org/abs/2512.15948
I welcome any feedback on these preliminary ideas.

We gave AI control of a real business YouTube video by Anthropic

important, and interesting!
youtu.be/5KTHvKCrQ00?...

I went over the task with chatgpt ('need a quote for seminar', etc). It first complained that there is no Act VI in Hamlet (fair), then when I pushed back, said

"Ah — that’s a lovely bit of trickster energy you’re bringing to your seminar!"

which made me want to do a Romeo-and-Juliette on myself

Cells can hunt prey, solve mazes, and learn from past experience. Does this mean they can think, at some level? Find out more at our upcoming session on "Cell Behavior and Cognition" at the ASCB/EMBO #cellbio2025 meeting, Saturday 1-3 pm rm 115.

Thank you for having me on BrainInspired, Paul @braininspired.bsky.social! It was such an honor to be on my favorite show—a rare place where we can leisurely talk about manifolds, latent circuits, power laws, and other esoteric ideas, and still be taken seriously in knowing they are all real.

On the Surprising Effectiveness of Large Learning Rates under Standard Width Scaling Scaling limits, such as infinite-width limits, serve as promising theoretical tools to study large-scale models. However, it is widely believed that existing infinite-width theory does not faithfully ...

📄 Paper: arxiv.org/abs/2505.22491

Catch our Spotlight at #NeurIPS2025 Today!

📅 Wed Dec 3 🕟 4:30 - 7:30 PM 📍 Exhibit Hall C,D,E — Poster #3903
Huge thanks to my amazing collaborators: @mohaas.bsky.social @sbordt.bsky.social @ulrikeluxburg.bsky.social

This is such an important point often lost in discussions about neuroimaging (e.g., no-report paradigms) or artificial intelligence (e.g., purported classification of, say, mental states or personality types). Any indirect measures tracking subjective states *depend* on subjective reports.

Overview of OMiSO

I'm excited to share our new #NeurIPS2025 paper with
@jsoldadomagraner @yugroupcmu @smithlabneuro!
We introduce OMiSO, a new adaptive brain stimulation framework with two technical advancements from our previous work MiSO.
Poster Session 1, Dec 3 11-2pm, Exhibit Hall C,D,E #2008
[1/n]

Memory traces bias new learning for hippocampal generalization The ability to use generalized prior experience to guide behavior in novel situations is a fundamental cognitive function[1][1]. While recent evidence suggests that the hippocampus supports generaliza...

www.biorxiv.org/content/10.1...