On Proper Learnability between Average- and Worst-case Robustness
원문 초록
Recently, Montasser et al. [2019] showed that finite VC dimension is not sufficient for proper adversarially robust PAC learning. In light of this hardness, there is a growing effort to study what type of relaxations to the adversarially robust PAC learning setup can enable proper learnability. In this work, we initiate the study of proper learning under relaxations of the worst-case robust loss. We give a family of robust loss relaxations under which VC classes are properly PAC learnable with sample complexity close to what one would require in the standard PAC learning setup. On the other hand, we show that for an existing and natural relaxation of the worst-case robust loss, finite VC dimension is not sufficient for proper learning. Lastly, we give new generalization guarantees for the adversarially robust empirical risk minimizer.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Lean Formalization of Generalization Error Bound by Rademacher Complexity
Prospective Learning: Learning for a Dynamic Future
Optimistic Rates for Learning from Label Proportions
Towards a theory of model distillation
SAT-Based PAC Learning of Description Logic Concepts
Privacy Induces Robustness: Information-Computation Gaps and Sparse Mean Estimation
토론
이 논문과 연결된 포럼 글이 아직 없습니다.