Borsuk-Ulam and Replicable Learning of Large-Margin Halfspaces

Recent advances in learning theory have established that, for total concepts, list replicability, global stability, differentially private (DP) learnability, and shared-randomness replicability coincide precisely with the finiteness of the Littlestone dimension. Does the same hold for partial concept classes? We answer this question by studying the large-margin half-spaces class, which has bounded Littlestone dimension and is purely DP-learnable and shared-randomness replicable even in high dimensions. We prove that the list replicability number of $\gamma$-margin half-spaces satisfies \[ \frac{d}{2} + 1 \le \mathrm{LR}(H_{\gamma}^d) \le d, \] which increases with the dimension $d$. This reveals a surprising separation for partial concepts: list replicability and global stability do not follow from bounded Littlestone dimension, DP-learnability, or shared-randomness replicability. By applying our main theorem, we also answer the following open problems. - We prove that any disambiguation of an infinite-dimensional large-margin half-space to a total concept class has unbounded Littlestone dimension, answering an open question of Alon et al. (FOCS '21). - We prove that the maximum list-replicability number of any *finite* set of points and homogeneous half-spaces in $d$-dimensional Euclidean space is $d$, resolving a problem of Chase et al. (FOCS '23). - We prove that any disambiguation of the Gap Hamming Distance problem in the large gap regime has unbounded public-coin randomized communication complexity. This answers an open problem of Fang et al. (STOC '25). We prove the lower bound via a topological argument involving the local Borsuk-Ulam theorem of Chase et al. (STOC '24). For the upper bound, we design a learning rule that relies on certain triangulations of the cross-polytope and recent results on the generalization properties of SVM.

Stability and List-Replicability for Agnostic Learners

Two seminal papers--Alon, Livni, Malliaris, Moran (STOC 2019) and Bun, Livni, and Moran (FOCS 2020)--established the equivalence between online learnability and globally stable PAC learnability in binary classification. However, Chase, Chornomaz, Moran, and Yehudayoff (STOC 2024) recently showed that this equivalence does not hold in the agnostic setting. Specifically, they proved that in the agnostic setting, only finite hypothesis classes are globally stable learnable. Therefore, agnostic global stability is too restrictive to capture interesting hypothesis classes. To address this limitation, Chase \emph{et al.} introduced two relaxations of agnostic global stability. In this paper, we characterize the classes that are learnable under their proposed relaxed conditions, resolving the two open problems raised in their work. First, we prove that in the setting where the stability parameter can depend on the excess error (the gap between the learner's error and the best achievable error by the hypothesis class), agnostic stability is fully characterized by the Littlestone dimension. Consequently, as in the realizable case, this form of learnability is equivalent to online learnability. As part of the proof of this theorem, we strengthen the celebrated result of Bun et al. by showing that classes with infinite Littlestone dimension are not stably PAC learnable, even if we allow the stability parameter to depend on the excess error. For the second relaxation proposed by Chase et al., we prove that only finite hypothesis classes are globally stable learnable even if we restrict the agnostic setting to distributions with small population loss.

Online Learning and Disambiguations of Partial Concept Classes

In a recent article, Alon, Hanneke, Holzman, and Moran (FOCS '21) introduced a unifying framework to study the learnability of classes of partial concepts. One of the central questions studied in their work is whether the learnability of a partial concept class is always inherited from the learnability of some ``extension'' of it to a total concept class. They showed this is not the case for PAC learning but left the problem open for the stronger notion of online learnability. We resolve this problem by constructing a class of partial concepts that is online learnable, but no extension of it to a class of total concepts is online learnable (or even PAC learnable).