Robust Prediction when Features are Missing
This addresses the issue of missing features in prediction tasks, which is a common problem in real-world applications, but the approach appears incremental as it builds on existing oracle-based methods.
The paper tackles the problem of making robust predictions when features are missing at test time, by developing an approach based on an oracle predictor's optimality properties, and demonstrates its robustness on real and synthetic data.
Predictors are learned using past training data which may contain features that are unavailable at the time of prediction. We develop an approach that is robust against outlying missing features, based on the optimality properties of an oracle predictor which observes them. The robustness properties of the approach are demonstrated on both real and synthetic data.