Mode Estimation with Partial Feedback
This work addresses the need for new theoretical frameworks in weakly supervised and active learning pipelines, though it appears incremental as it adapts existing methods like bandit algorithms and entropy coding to a specific setting.
The paper tackles the problem of estimating the mode of a distribution using partial feedback, developing a solution that is statistically and computationally efficient.
The combination of lightly supervised pre-training and online fine-tuning has played a key role in recent AI developments. These new learning pipelines call for new theoretical frameworks. In this paper, we formalize core aspects of weakly supervised and active learning with a simple problem: the estimation of the mode of a distribution using partial feedback. We show how entropy coding allows for optimal information acquisition from partial feedback, develop coarse sufficient statistics for mode identification, and adapt bandit algorithms to our new setting. Finally, we combine those contributions into a statistically and computationally efficient solution to our problem.