Adaptive Sampling for Stochastic Risk-Averse Learning
This addresses the need for robust performance in high-stakes machine learning applications, offering a novel method for risk-averse optimization.
The paper tackles the problem of training models in a risk-averse manner by optimizing the Conditional Value-at-Risk (CVaR) to improve performance on difficult examples, proposing an adaptive sampling algorithm that uses distributionally robust formulation and Determinantal Point Processes (DPPs) for efficient scaling, and demonstrates effectiveness on large-scale convex and non-convex tasks.
In high-stakes machine learning applications, it is crucial to not only perform well on average, but also when restricted to difficult examples. To address this, we consider the problem of training models in a risk-averse manner. We propose an adaptive sampling algorithm for stochastically optimizing the Conditional Value-at-Risk (CVaR) of a loss distribution, which measures its performance on the $α$ fraction of most difficult examples. We use a distributionally robust formulation of the CVaR to phrase the problem as a zero-sum game between two players, and solve it efficiently using regret minimization. Our approach relies on sampling from structured Determinantal Point Processes (DPPs), which enables scaling it to large data sets. Finally, we empirically demonstrate its effectiveness on large-scale convex and non-convex learning tasks.