Balanced Linear Contextual Bandits
This addresses the challenge of robust learning in contextual bandits for applications like recommendation systems, though it is incremental as it builds on existing balancing techniques.
The authors tackled the problem of estimation bias in linear contextual bandits by integrating balancing methods from causal inference, resulting in algorithms that match state-of-the-art theoretical regret bounds and show strong practical advantages on supervised learning datasets and synthetic examples.
Contextual bandit algorithms are sensitive to the estimation method of the outcome model as well as the exploration method used, particularly in the presence of rich heterogeneity or complex outcome models, which can lead to difficult estimation problems along the path of learning. We develop algorithms for contextual bandits with linear payoffs that integrate balancing methods from the causal inference literature in their estimation to make it less prone to problems of estimation bias. We provide the first regret bound analyses for linear contextual bandits with balancing and show that our algorithms match the state of the art theoretical guarantees. We demonstrate the strong practical advantage of balanced contextual bandits on a large number of supervised learning datasets and on a synthetic example that simulates model misspecification and prejudice in the initial training data.