arXiv Machine Learning
Jun 29
Surprises in Proper Positive-Only Learning
arXiv:2606. 28309v1 Announce Type: cross Abstract: Binary classification from positive-only samples is a variant of PAC learning in which the learner receives i.
Combinatorial Characterizations of the Learnability of Conditional Distributions
arXiv:2610. 31890v1 Announce Type: new Abstract: We study the PAC-learnability of function classes that map inputs to probability distributions over a finite space.
Related stories
arXiv Machine Learning
Jul 2
Recursively Enumerably Representable Classes and Computable Versions of the Fundamental Theorem of Statistical Learning
arXiv:2511. 02644v2 Announce Type: replace Abstract: We study computable probably approximately correct (CPAC) learning, where learners are required to be computable functions.
arXiv Machine Learning
Sep 1
A Borel Concept Class of VC Dimension One with a Non-PAC Consistent Learner in ZFC
arXiv:2608.30246v1 Announce Type: cross Abstract: The fundamental theorem of statistical learning states that, under suitable measurability assumptions, finite Vapnik--Chervonenkis (VC) dimension gua...
arXiv Machine Learning
6d ago
Actually, All Learnable Distribution Classes Are Privately Learnable
arXiv:2610. 28878v1 Announce Type: new Abstract: We show that every class of distributions that is learnable up to arbitrarily small target error in total variation distance with finitely many samples is also learnable under ($\varepsilon, \delta$)-differential privacy up to arbitrarily small target error with finitely many samples.
arXiv Machine Learning
Aug 7
Realizable Bayes-Consistency for General Metric Losses
arXiv:2605. 03823v3 Announce Type: replace Abstract: We study strong universal Bayes-consistency in the realizable setting for learning with general metric losses, extending classical characterizations beyond $0$-$1$ classification (Bousquet et al.
arXiv Machine Learning
Jun 24
Relatively Smart: A New Approach for Instance-Optimal Learning
arXiv:2603. 01346v2 Announce Type: replace Abstract: We revisit the framework of Smart PAC learning, which seeks supervised learners which compete with semi-supervised learners that are provided full knowledge of the marginal distribution on unlabeled data.
arXiv Machine Learning
Oct 8
Boosting and the Expressive Power of Simple Weak Learners via the $\gamma$-VC Dimension
arXiv:2610. 10383v1 Announce Type: new Abstract: Boosting converts weak hypotheses with a small edge over random guessing into highly accurate predictors, but the expressive power of the resulting classifier can depend strongly on the structure of the base class.
arXiv Machine Learning
Oct 2
Scale-Sensitive Shattering: Learnability and Evaluability at Optimal Scale
arXiv:2605.13684v2 Announce Type: replace Abstract: We study the optimal scale at which real-valued function classes exhibit uniform convergence and learnability. Our main result establishes a scale-...
arXiv Machine Learning
Jun 2
How Many Domains Suffice for Domain Generalization? A Tight Characterization via the Domain Shattering Dimension
arXiv:2506. 16704v3 Announce Type: replace Abstract: We study a fundamental question of domain generalization: given a family of domains (i.
arXiv Machine Learning
Jul 1
A Complete Characterization of Learnability for Adversarial Noisy Bandits
arXiv:2605. 09200v2 Announce Type: replace Abstract: We study adversarial noisy bandits given a known function class $\mathcal{F}$.
arXiv Machine Learning
Aug 27
Two Dimensions Govern Agnostic Multiclass Transductive Learning
arXiv:2608. 25326v1 Announce Type: new Abstract: In transductive classification, an adversary fixes a labeled population, one label is hidden uniformly, and the learner sees all remaining labels.
arXiv Machine Learning
Aug 28
Algorithmic Principles For Multiclass Learning Are Hard To Come By: Limits of Regularization and Proper Learning
The paper investigates fundamental limits of algorithmic principles in multiclass learning, specifically proper learning and regularization. It shows that learning cannot always be reduced to proper learning even with an enlarged hypothesis class, that proper learners may need a sublinear number of errors that can be arbitrarily large, and that regularization (SRM or local) is not universally sufficient. The authors also provide a positive theory giving sufficient conditions for SRM learnability and a characterization via integrability of revealed preferences.