Adaptive Sampling for Stochastic Risk-Averse Learning

Abstract

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 $\alpha$ 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.

References in corpus (17)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Improving neural networks by preventing co-adaptation of feature detectors
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• Optimization, Learning, and Games with Predictable Sequences
• Risk-Aversion in Multi-armed Bandits
• Learning with Average Top-k Loss
• Minimizing the Maximal Loss: How and Why?
• Distributionally Robust Bayesian Optimization
• Fairness risk measures
• Statistical Learning with Conditional Value at Risk
• Distributionally Robust Optimization and Generalization in Kernel Methods
• Learning Bounds for Risk-sensitive Learning
• PAC-Bayesian Bound for the Conditional Value at Risk
• Exact sampling of determinantal point processes with sublinear time preprocessing
• Online Variance Reduction with Mixtures
• Learning in Non-convex Games with an Optimization Oracle
• Fast Mixing Markov Chains for Strongly Rayleigh Measures, DPPs, and Constrained Sampling

Cited by in corpus (6)

• Large-Scale Methods for Distributionally Robust Optimization
• Adaptive sampling strategies for risk-averse stochastic optimization with constraints
• Federated Learning with Superquantile Aggregation for Heterogeneous Data
• Device Heterogeneity in Federated Learning: A Superquantile Approach
• Noisy Linear Convergence of Stochastic Gradient Descent for CV@R Statistical Learning under Polyak-Łojasiewicz Conditions
• A Survey of Learning Criteria Going Beyond the Usual Risk