Pith. sign in

REVIEW 2 cited by

Adaptive and oblivious statistical adversaries are equivalent

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.13548 v2 pith:WFXGBP5Q submitted 2024-10-17 cs.LG cs.CCcs.DS

classification cs.LGcs.CCcs.DS
keywords adversariessamplewhenadversarycorruptssample-adaptivesample-obliviousstatistical
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

We resolve a fundamental question about the ability to perform a statistical task, such as learning, when an adversary corrupts the sample. Such adversaries are specified by the types of corruption they can make and their level of knowledge about the sample. The latter distinguishes between sample-adaptive adversaries which know the contents of the sample when choosing the corruption, and sample-oblivious adversaries, which do not. We prove that for all types of corruptions, sample-adaptive and sample-oblivious adversaries are \emph{equivalent} up to polynomial factors in the sample size. This resolves the main open question introduced by [BLMT22] and further explored in [CHL+23]. Specifically, consider any algorithm $A$ that solves a statistical task even when a sample-oblivious adversary corrupts its input. We show that there is an algorithm $A'$ that solves the same task when the corresponding sample-adaptive adversary corrupts its input. The construction of $A'$ is simple and maintains the computational efficiency of $A$: It requests a polynomially larger sample than $A$ uses and then runs $A$ on a uniformly random subsample.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Agnostic Learning under Targeted Poisoning: Optimal Rates and the Role of Randomness

    cs.LG 2025-06 conditional novelty 8.0 of 10

    The optimal excess error for agnostic learning under instance-targeted poisoning is eTheta(sqrt(d eta)), achieved by a randomized learner and unavoidable even against adversaries who see the learner's random seed.

  2. On the Learnability of Distribution Classes with Adaptive Adversaries

    cs.LG 2025-09 conditional novelty 6.0 of 10

    A realizably learnable distribution class is proven not learnable against an adaptive additive adversary, while the same class is learnable against oblivious additive tampering.

Pith tools