regret scaling

A concept in online learning relating to the trade-off between learning speed and the cumulative regret (or difference between predicted and actual outcomes) over time.

5 papers

Comparing Uniform Price and Discriminatory Multi-Unit Auctions through Regret Minimization
Online Learning of Pure States is as Hard as Mixed States
Prediction with expert advice under additive noise
Reward-oriented Causal Representation Learning
Stochastic Shortest Path with Sparse Adversarial Costs