Machine learning theory

Adaptive Exploration for Latent-State Bandits
We propose adaptive, state-model-free bandit algorithms for latent-state (confounded, non-stationary) environments.
Jikai Jin, Kenneth Hung, Sanath Kumar Krishnamurthy, Baoyi Shi, Congshan Zhang
Cite DOI ArXiv
Policy Learning with Abstention
We study policy learning with the ability to abstain, and demonstrate its usefulness in ensuring safety.
Ayush Sawarni, Jikai Jin, Justin Whitehouse, Vasilis Syrgkanis
Cite ArXiv
Sharp Structure-Agnostic Lower Bounds for General Linear Functional Estimation
Sharp minimax theory for structure-agnostic estimation of general linear functionals, including optimality of first-order debiasing and doubly robust estimators.
Jikai Jin, Vasilis Syrgkanis
Cite DOI ArXiv