A Theory of Optimistically Universal Online Learnability for General Concept Classes
This work addresses foundational theoretical questions in online learning for all concept classes, establishing equivalence between minimal assumptions and optimistically universal learnability.
The paper fully characterizes concept classes that are optimistically universally online learnable with binary labels, resolving minimal data process assumptions and designing general learning algorithms for both realizable and agnostic cases.
We provide a full characterization of the concept classes that are optimistically universally online learnable with $\{0, 1\}$ labels. The notion of optimistically universal online learning was defined in [Hanneke, 2021] in order to understand learnability under minimal assumptions. In this paper, following the philosophy behind that work, we investigate two questions, namely, for every concept class: (1) What are the minimal assumptions on the data process admitting online learnability? (2) Is there a learning algorithm which succeeds under every data process satisfying the minimal assumptions? Such an algorithm is said to be optimistically universal for the given concept class. We resolve both of these questions for all concept classes, and moreover, as part of our solution, we design general learning algorithms for each case. Finally, we extend these algorithms and results to the agnostic case, showing an equivalence between the minimal assumptions on the data process for learnability in the agnostic and realizable cases, for every concept class, as well as the equivalence of optimistically universal learnability.