Generalization Bounds and Consistency for Latent Structural Probit and Ramp Loss

Part of Advances in Neural Information Processing Systems 24 (NIPS 2011)

Bibtex Metadata Paper


Joseph Keshet, David McAllester


We consider latent structural versions of probit loss and ramp loss. We show that these surrogate loss functions are consistent in the strong sense that for any feature map (finite or infinite dimensional) they yield predictors approaching the infimum task loss achievable by any linear predictor over the given features. We also give finite sample generalization bounds (convergence rates) for these loss functions. These bounds suggest that probit loss converges more rapidly. However, ramp loss is more easily optimized and may ultimately be more practical.