Liberal Arts and Sciences
Bhattacharya Lab
Led by Sohom Bhattacharya · Statistics
What we study
The Bhattacharya Lab studies high-dimensional statistics and machine learning, focusing on data heterogeneity and complex dependencies. They use computational methods to analyze statistical models and causal inference.
What working here is like
- Computational
- No organisms (data only)
- Coding required
- Projects
- 2
- Papers
- 5
Current projects
Multi-task and transfer learning theory
How well learning methods generalize when data come from several related tasks.
Causal inference under network interference
Estimating cause and effect when one unit's treatment affects its neighbours.
Summaries are written by AI in plain language from public symposium abstracts and may be imperfect.
Recent papers, explained
- Inferences on Mixing Probabilities and Ranking in Mixed-Membership Models
2026 · Journal of the American Statistical Association
Develops a method to estimate uncertainty in mixed-membership network models.
- Gibbs measures with multilinear forms
2026 · The Annals of Applied Probability
Studies Gibbs measures with multilinear forms and shows how replica symmetry affects their behavior.
Show 3 more papersShow fewer
- Nonsense associations in Markov random fields with pairwise dependence
2025 · Biometrika
Examines nonsense correlations in Markov random fields caused by pairwise dependence.
- Sharp Signal Detection Under Ferromagnetic Ising Models
2025 · IEEE Transactions on Information Theory
Analyzes signal detection in ferromagnetic Ising models to find sharp detection boundaries.
- Causal effect estimation under network interference with mean-field methods
2025 · The Annals of Statistics
Creates algorithms to estimate causal effects in networks with interference.
Plain-language summaries are AI-written and reviewed by us; check the paper for details.
How to reach out
Sohom Bhattacharya hasn't set up this page yet. You can also look them up in the UF directory (search “Sohom Bhattacharya”).