Raaz RSK Dwivedi
 |
Assistant Professor, Cornell Tech, Cornell University! My research interests are in the areas of causal inference, reinforcement learning, and distribution compression with applications in personalized decision-making and healthcare. If you are a PhD student or a postdoc at Cornell and interested in these topics, feel free to email me. I am also actively looking to recruit motivated and talented students applying to Cornell (see here). |
News
Upcoming presentations:
Mar 4, 4:10-5:10 PM: Statistics Seminar, Columbia!
Apr 18, 4:15-5:15 PM: Operations Research Seminar, MIT!
May 7, 11:30 AM-12:30 AM ET: Online Causal Inference Seminar!
Jun 28, 11:15 AM-12:00 PM ET: Tom Ten Have Symposium on Mental Health Statistics, Weil Cornell!
July 18-19: Mini Workshop on Individualized Decisions, Simons/UC Berkeley!
Aug 7, 10:30 AM-12:00 PM ET Joint Statistical Meeting (JSM), Portland!
Aug 13, 4:20-6:00 PM ET, Parallel 1, Room 196: ESIF Economics and AI+ML Meeting, Cornell, Ithaca!
Oct 20-23: Informs, Seattle!
Recent preprints and publications:
May 2024: Debiased Distribution Compression, joint work with Lingxiao Li and Lester Mackey accepted at ICML 2024!
Apr 2024: Kernel Thinning, joint work with Lester Mackey accepted at JMLR!
Apr 2024: FairPair: A Robust Evaluation of Biases in Language Models through Paired Perturbations, joint work with Jane Dwivedi-Yu and Timo Schick up on arxiv!
Mar 2024: ‘‘Did we personalize? Assessing personalization by an online reinforcement learning algorithm using resampling’’, joint work with Susobhan Ghosh, Raphael Kim, Prasidh Chhabria, Predrag Klasjna, Peng Liao, Kelly Zhang, and Susan Murphy, accepted in Machine Learning Journal!
Feb 2024: ‘‘Doubly Robust Inference in Causal Latent Factor Models’’, joint work with Alberto Abadie, Anish Agarwal, and Abhin Shah up on arxiv!
Selected presentations
Integrating Double Robustness Into Causal Latent Factor Models || TALK Video || slides
2024: Columbia Stats, MIT ORC, Online Causal Inference Seminar, Cornell Weil (Tom Ten Have)
From HeartSteps To HeartBeats: Personalized Decison-making || TALK Video || animated slides | slides ||
2023: Cornell ORIE Industry Summit, AI CS, Stats & DS, CAM, Yale Stats & DS, Gatsby UCL, Stanford GSB OIT Group, UW Madison Stats, UIUC CS, Wharton Stats, UCLA Stats; 2022: Cornell ORIE, Cornell Tech ORIE, Rutgers Stats, Stanford EE (ISL), UC Berkeley EECS (BLISS)
Counterfactual Inference
In Sequential Experiments || preprint 1, preprint 2 | slides | poster ||
With Unobserved Confounding || preprint 3 | slides | poster ||
2023: ACM-FCRC, Informs APS, NESS, ACIC, MIT, IIM-A; 2022: NeurIPS, Cornell YRW, RSS, Informs, IMS, ACIC, SCM Princeton, Simons, MIT, Harvard
Distribution Compression
For Integration: Kernel Thinning And Compress || TALK Video || COLT (KT), ICLR 1 (GKT), ICLR 2 (Compress) | slides ||
For Hypothesis Testing: Compress Then Test || AISTATS (CTT) | slides | poster ||
2024: KAUST, 2023: JSM, MCM, AISTATS, MIT; 2022: RSS, SIAM, MSRI, JSM, ICLR, AABI, Harvard
Selected Other Topics/Talks
Non-asymptotic Guarantees For High-Dimensional Sampling || JMLR (MALA) | JMLR (HMC) | JMLR (Vaidya/John walks) | slides ||
Stable Discovery Of Interpretable Subgroups In Causal Studies || ISR | slides | poster ||
Revisiting Minimum Description Length Complexity In Overparameterized Models || preprint | slides | poster | video ||
Brief Bio
2021-2023: FODSI Postdoctoral Fellow, Harvard & MIT, co-advised by Prof. Susan Murphy & Prof. Devavrat Shah
2015-2021: Ph.D., EECS, UC Berkeley, co-advised by Prof. Martin Wainwright & Prof. Bin Yu (thesis)
2010-2014: B. Tech., EE, IIT Bombay, advised by Prof. Vivek Borkar (thesis)
Contact: Rhodes 226 Ithaca, dwivedi@cornell.edu, Google Scholar
Fun Stat
Website visitor counts by country since Apr 3, 2022: