Excited to share our new work on dynamics in compositional RL with curricula accepted ICML Conference 2024! 📄 [arxiv.org/abs/2402.18361] 👯 Team = me, Stefano Sarao Mannelli - Hiring! Andrew Saxe TL;DR We propose a simple compositional RL model amenable to exact learning dynamics analysis which can account

👋 Attending #ICML2024 in 🇦🇹 next week? Don't forget to check out work involving our researchers!

It’s today ICML Conference 25.07 Thursday 11.30-13.00 #1407 Come and chat! w/Stefano Sarao Mannelli - Hiring! Andrew Saxe

Come join our team in London as a research assistant! Rodent decision making, miniscopes and hippocampus! Ideal for a recent graduate wanting lab experience before PhD. UCL is an amazing place to do neuroscience - come and find out! ucl.ac.uk/work-at-ucl/se…

Interested in exactly solvable models of learning dynamics and implicit bias? Come check out our "Get Rich Quick" poster at the HiLD Workshop ICML Conference at 10am! With Daniel Kunin, myself, Clémentine Dominé 🍊, FENG CHEN, David Klindt, Andrew Saxe, and Surya Ganguli.

Only 3 days left to apply to join my group Chalmers University of Technology as a PhD student: deadline July 31st. Work with me to uncover the mechanisms behind bias generation and amplification in ML! 🔗 chalmers.se/en/about-chalm…

Yesterday at #CoLLAs2024, the talented Jin Lee and I gave a tutorial on 'Advances in Continual Learning Theory'. It's so exciting to see the growing interest in continual learning! Jin Hwa Lee CoLLAs 2025

Super excited to give a talk on this project at CogCompNeuro #CCN2024 on Friday morning. Come chat all things sleep, regularisation and insight at my poster C165 afterwards :)

Today is my first time #CCN2024. I'm very excited and mildly jetlagged! I'll be presenting two works: 📃B51 On compositional generalisation and curriculum learning; 📃C93 A framework for modelling critical learning. Looking forward to some feedback and follow-up ideas! 👇

Why Do Animals Need Shaping? A Theory of Task Composition and Curriculum Learning (B51) 🔗 2024.ccneuro.org/poster/?id=191 📄arxiv.org/abs/2402.18361 👯‍♂️ Jin Hwa Lee, me, Andrew Saxe 📅 Thu 8 Aug 1:30 p.m.

A Minimal and Flexible Model for Investigating Critical Learning Periods (C96) 🔗 2024.ccneuro.org/poster/?id=182 👯‍♂️ Sebastian Lee, me, Andrew Saxe 📅 Fri 9 Aug 11:15 a.m.

Come to poster C146 CogCompNeuro at 11:15-1:15 today for some evidence showing how learning shapes cognitive control allocation! 2024.ccneuro.org/poster/?id=136

We started today the new edition of Analytical Connectionism! This week Linda Smith, Cengiz Pehlevan and Jonathan Cohen will give us in-depth lectures on topics ranging from development to analytical methods for NNs! Happy days! 🎈🎈🎈

Really excited to share that I'm hiring for a Research Scientist position in our team! If you're interested in the kind of cognitively-oriented work we've been doing on learning & generalization, data properties, representations, LMs, or agents, please check it out!

Our paper on the effect of stimulus temporal regularity on visual decisions is out in its final form: nature.com/articles/s4146… Thanks to reviewers whose comments made it a better paper! Matthias Fritsche Antara Majumdar 🧠👩🏽‍🔬💫 Samuel Liebana Lauren Strickland Rafal Bogacz.

Less than 10 days left to apply for the first postdoc position in my lab! 🌱 If you're interested in understanding ML at a fundamental level using statistical physics and probability tools, here is the link 🔗chalmers.se/en/about-chalm… Deadline for applications: August 31st

What an amazing crowd of young researchers at the Analytical Connectionism Summer School hackathon! The past two weeks have been so much fun. 🙌 SueYeon Chung Stefano Sarao Mannelli Andrew Saxe #ACSS2024

To ask how similar the brain is to a neural network we need a similarity metric. In a new paper I asked how much the metric matters to downstream conclusions, and, upshot, it matters a great deal. biorxiv.org/content/10.110… (1/7)

What is the nature of sensory prediction error signals in the neocortex? Our research published in nature reveals that these error signals are amplified responses to unexpected inputs, mediated by a cooperative thalamocortical circuit mechanism! 1/4 nature.com/articles/s4158…