Mehryar Mohri

Next-Token Prediction and Regret Minimization

High-Dimensional Calibration from Swap Regret

Improved Balanced Classification with Theoretically Grounded Loss Functions

Principled Algorithms for Optimizing Generalized Metrics in Binary Classification

Balancing the Scales: A Theoretical and Algorithmic Framework for Learning from Imbalanced Data

Principled Model Routing for Unknown Mixtures of Source Domains

Mastering Multiple-Expert Routing: Realizable $H$-Consistency and Strong Guarantees for Learning to Defer