Active Online Learning with Hidden Shifting Domains
This work addresses the challenge of efficient online adaptation for machine learning systems in dynamic environments where labeling is costly, offering an incremental improvement over existing query strategies.
The paper tackles the problem of online learning with domain shifts and expensive labeling by proposing an algorithm that adaptively balances regret and label queries for hidden shifting domains. It achieves lower regret than uniform and greedy querying methods with equal labeling budgets, as demonstrated on synthetic and realistic datasets.
Online machine learning systems need to adapt to domain shifts. Meanwhile, acquiring label at every timestep is expensive. We propose a surprisingly simple algorithm that adaptively balances its regret and its number of label queries in settings where the data streams are from a mixture of hidden domains. For online linear regression with oblivious adversaries, we provide a tight tradeoff that depends on the durations and dimensionalities of the hidden domains. Our algorithm can adaptively deal with interleaving spans of inputs from different domains. We also generalize our results to non-linear regression for hypothesis classes with bounded eluder dimension and adaptive adversaries. Experiments on synthetic and realistic datasets demonstrate that our algorithm achieves lower regret than uniform queries and greedy queries with equal labeling budget.