An Axiomatic Approach to Loss Aggregation and an Adapted Aggregating Algorithm
This work provides a theoretical extension for online learning algorithms, but it is incremental as it adapts existing methods to more general aggregation functions without broad practical validation.
The paper tackles the problem of generalizing loss aggregation functions in online learning under expert advice, characterizing reasonable functions as quasi-sums and proposing a variant of the Aggregating Algorithm that maintains theoretical properties like a time-independent bound on quasi-sum regret.
Supervised learning has gone beyond the expected risk minimization framework. Central to most of these developments is the introduction of more general aggregation functions for losses incurred by the learner. In this paper, we turn towards online learning under expert advice. Via easily justified assumptions we characterize a set of reasonable loss aggregation functions as quasi-sums. Based upon this insight, we suggest a variant of the Aggregating Algorithm tailored to these more general aggregation functions. This variant inherits most of the nice theoretical properties of the AA, such as recovery of Bayes' updating and a time-independent bound on quasi-sum regret. Finally, we argue that generalized aggregations express the attitude of the learner towards losses.