ᴘᴇᴛᴇʀ ɢ. ᴄʜᴀɴɢ
@petergchang
Followers
257
Following
2K
Media
1
Statuses
35
PhD Student | @MIT EECS (with @m_sendhil) | AI for Scientific Discovery
Cambridge, MA
Joined May 2022
I’m hiring a pre-doc! Come work with me on how AI is changing the labor market and how algorithms impact markets. Non-econ backgrounds welcome. Application details below – excited to collaborate! Start: Summer 2026 Deadline: Nov 1, 2025 https://t.co/2joGp5czWN
@predoc_org
17
92
396
Can #LLMs grasp the real world? MIT & Harvard researchers (@m_sendhil, @asheshrambachan, @petergchang, @keyonV) propose a new way to test how predictive AI applies knowledge across domains. Learn more: https://t.co/npsSXgyHyT
0
5
5
A common takeaway from "the bitter lesson" is we don't need to put effort into encoding inductive biases, we just need compute. Nothing could be further from the truth! Better inductive biases mean better scaling exponents, which means exponential improvements with computation.
20
34
420
If you’re at ICML, please come check out the World Models workshop I am co-organizing tomorrow!
Join us for the Workshop on Assessing World Models at ICML tomorrow! When: Friday July 17, 8:45am-5:15pm Where: West Ballroom B (same floor as registration)
1
1
6
Researchers from Harvard, Keyon Vafa (@keyonV) and MIT, Peter Chang (@petergchang), Ashesh Rambachan (@asheshrambachan), and Sendhil Mullainathan (@m_sendhil) have published what I consider the most interesting study on the abilities of AI models in 2025. They wanted to address
11
27
79
The first paper I’ve worked on as a PhD student is out! Very proud of this work.
Can an AI model predict perfectly and still have a terrible world model? What would that even mean? Our new ICML paper formalizes these questions One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵
1
3
22
& here are some of my favorite papers on the unification of flows and diffusion. What a decade!! (From my presentation here: https://t.co/B93hcTWPFy)
How Diffusion unification went: > score based model > then DDPM came along > we have two formalism, DDPM & SBM > SDE came to unify them > now we have Score, DDPM & SDE > Then came flow matching to unify them > now we have Score, DDPM, SDE & Flow Models > Then consistency models
5
41
316
Excited to share our latest story! We found disentangled memory representations in the hippocampus that generalized across time and environments, despite the seemingly random drift and remapping of single cells. This code enabled the transfer of prior knowledge to solve new tasks
16
163
1K
Worked on a little project which helps converting PyTorch models to JAX PyTrees (e.g. for usage in Equinox). You can also visualise both networks using the excellent Penzai library! In the video, I'm converting Resnet18 to a PyTree. https://t.co/kwRn7OFyen
0
20
137
I was so lucky to be able to have Danny Kahneman as a best friend and collaborator for decades. He usually ended our conversations with "to be continued..." but I now have to simulate his part which is impossible. My favorite image of us "working".
105
1K
9K
This looks cool: Mathieu Blondel (@mblondel_ml) and Vincent Roulet have posted a first draft of their book on arXiv:
arxiv.org
Artificial intelligence has recently experienced remarkable advances, fueled by large models, vast datasets, accelerated hardware, and, last but not least, the transformative power of...
4
79
371
If you know Torch, I think you can code for GPU now with OpenAI's Triton language. We made some puzzles to help you rewire your brain. Starts easy, but gets quickly to fun modern models like FlashAttention and GPT-Q. Good luck! https://t.co/psu31KpdR6
17
197
1K
Can AI Solve Science? (podcast version with Q&A) now available on YouTube: https://t.co/8c6DIo3eSa
12
52
234
My friend Chenyang (twitterless) has written a nice tutorial on diffusion models, from the "projection onto manifold" perspective. My favorite part is the *extremely simple* pedagogical codebase, which I've been using myself for quick experiments. See: https://t.co/iLCXOu2qhj
4
138
802
I've been wondering: can AI solve science? I just did some analysis... https://t.co/C6eymt4QhV
30
159
770
If you watched my "Getting Started with CUDA for Python Programmers", and are ready to go even faster, then this new video is for you! Fast CUDA isn't just about having all the threads go brrrr in parallel, but also give them fast memory to play with. https://t.co/PVw36HuV62
10
157
1K
(0/17) Grab your🍿 for a thread on some mysteries and explanations connecting flat minima, second order optimization, weight noise, gradient norm penalty, and activation functions😱 There is also a video presentation if you prefer:
5
72
450
Interested in working on Bayesian neural networks/ deep ensembles? Here's a reading list to get you started! I've taken away alot of the clutter and tried to include the main papers from most major groups and when the predictive task is classification. ⬇️ https://t.co/sIvpmikkfv
github.com
A primer on Bayesian Neural Networks. The aim of this reading list is to facilitate the entry of new researchers into the field of Bayesian Deep Learning, by providing an overview of key papers. Mo...
1
8
58
Announcing a new #JAX and #Equinox nonlinear optimisation library: ⭐️ Optimistix ⭐️ (GitHub: https://t.co/UqmZDnTyL1) - Minimisation - Nonlinear least-squares - Root-finding - Fixed-points With blazing fast compile times and fast run times 🔥 1/7
3
36
267
The trilogy is complete! My "Advanced Topics" book is officially released today. Buy it on Amazon, or get it for free at https://t.co/yWK8zMbMRA.
61
749
5K