Tried Frontdoor formula with chatgpt. It seems my job is still safe.

If we view physics as an endeavor to find the most efficient compression to summarize real-world data, the line between fields starts to blur.

Please check out our work "Causal Imitation for Markov Decision Processes: a Partial Identification Approach" (with Kangrui Ruan, Sharon (Colubmia) and Elias Bareinboim) at #NeurIPS2024! - Thu 12 Dec 4:30 pm — 7:30 pm - West Ballroom A-D #5002 - Link: neurips.cc/virtual/2024/p…

Welcome to check out our new work at ICML!

2/5 "Automatic Reward Shaping from Confounded Offline Data" (w/Mingxuan Li, Junzhe Zhang) Wed, 1 PM (East Ballroom, 1802) Link: causalai.net/r123.pdf Reward shaping is a popular technique for addressing the sparse reward problem in RL by injecting additional

4/5 (UAI) "Eligibility Traces for Confounding Robust Off-Policy Evaluation" (w/ Junzhe Zhang) Link: causalai.net/r105.pdf The main goal of off-policy evaluation is to enable agents to learn from observational data without deploying new policies. While current methods

5/5 Last but definitely not least, I’m honored to be giving a keynote on Wednesday (7/23) titled "Towards Causal Artificial Intelligence." For details, see: auai.org/uai2025/keynot… Here’s a short abstract: While many AI scientists and engineers believe we are on the verge of

Posting non-academic stuff. Usyk just proved Ali would have been unbeatable in today's era.

I mean, how did you bid on causal inference papers if you are not familiar with basic literature?

Looking forward to working with Theta Network.

ACs implementing the review protocols can frustrate some reviewers. You can literally read it through their replies and comments. Lol #NeurIPS

The sooner we recognize that LLMs' "implicit reasoning" is simply MCMC sampling or the Euler method approximating a continuous function, the easier it becomes to make progress on its inference.