Devavrat Shah
Andrew (1956) and Erna Viterbi Professor
Develops practical algorithms for statistical inference and social data processing, including sparse matrix estimation, collaborative filtering, and ranking.
Devavrat Shah's lab at MIT works on learning theory & optimization, efficient training & inference and human-AI collaboration & trust. 17 papers since 2023, 8 at top venues, mostly NeurIPS and ICML. 6 PhD students. 34 alumni. Applications close .
At a glance
- papers, 2023–2026
- 4/ year
- at T1 venues
- 47%
- PhD students
- 6
- MS students
- 0
- alumni listed
- 34
Papers by yearfrom DBLP, 2023–2026
- 2 papers2026 · T1 0 · collecting
- 2 papers2025 · T1 1
- 4 papers2024 · T1 1
- 9 papers2023 · T1 6
T1 venuesOtherVenues: NeurIPS, ICML, AAAI/ACM AIES, CDC, COLT
Research focuspaper titles that use each topic's words
- Learning theory & optimization2 papers
- Human-AI collaboration & trust2 papers
- filtering1 paper
- Efficient training & inference0 papers
- social data0 papers
Paper list
From DBLP.
20262 papers
20252 papers
20244 papers
- AAAI/ACM AIESA Causal Framework to Evaluate Racial Bias in Law Enforcement Systems (opens in a new tab)
- IEEE Trans. Inf. TheoryEstimation of Skill Distributions (opens in a new tab)
- NeurIPS (T1)Human Expertise in Algorithmic Prediction (opens in a new tab)
- IEEE Trans. Signal Inf. Process. over NetworksUnifying Epidemic Models With Mixtures (opens in a new tab)
20239 papers
- NeurIPS (T1)Auditing for Human Expertise (opens in a new tab)
- COLT (T1)Causal Matrix Completion (opens in a new tab)
- NSDI (T1)CausalSim: A Causal Framework for Unbiased Trace-Driven Simulation (opens in a new tab)
- ICML (T1)Counterfactual Identifiability of Bijective Causal Models (opens in a new tab)
- IEEE Trans. Inf. TheoryFederated Optimization of Smooth Loss Functions (opens in a new tab)
- CDCGradient Descent with Low-Rank Objective Functions (opens in a new tab)
- ICML (T1)Matrix Estimation for Individual Fairness (opens in a new tab)
- IEEE Trans. Inf. TheoryRobust Max Entrywise Error Bounds for Tensor Estimation From Sparse Observations via Similarity-Based Collaborative Filtering (opens in a new tab)
- NeurIPS (T1)SAMoSSA: Multivariate Singular Spectrum Analysis with Stochastic Autoregressive Noise (opens in a new tab)
Where alumni went34 listed
Apply to MITFall 2027
MIT EECS Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- See program page
- English
- TOEFL/IELTS
- Fee
- USD 90
Confirmed on the program page.
Similar labsnearest by research summary
- CNSJTU
Shengzhong Liu
- Distributed, cloud & edge systems
- GBUCL
Paolo Barucca
- Computational social science
- CHEPFL
Rachid Guerraoui
- Trustworthy & safe ML
- CNPeking
Hong Mei
- Graph learning
- GBCambridge
Nicholas Lane
- Efficient training & inference
- USGeorgia Tech
Francisco Romero
- Distributed, cloud & edge systems
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
MITcloses