ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1905.10360
20
29

The advantages of multiple classes for reducing overfitting from test set reuse

24 May 2019
Vitaly Feldman
Roy Frostig
Moritz Hardt
ArXivPDFHTML
Abstract

Excessive reuse of holdout data can lead to overfitting. However, there is little concrete evidence of significant overfitting due to holdout reuse in popular multiclass benchmarks today. Known results show that, in the worst-case, revealing the accuracy of kkk adaptively chosen classifiers on a data set of size nnn allows to create a classifier with bias of Θ(k/n)\Theta(\sqrt{k/n})Θ(k/n​) for any binary prediction problem. We show a new upper bound of O~(max⁡{klog⁡(n)/(mn),k/n})\tilde O(\max\{\sqrt{k\log(n)/(mn)},k/n\})O~(max{klog(n)/(mn)​,k/n}) on the worst-case bias that any attack can achieve in a prediction problem with mmm classes. Moreover, we present an efficient attack that achieve a bias of Ω(k/(m2n))\Omega(\sqrt{k/(m^2 n)})Ω(k/(m2n)​) and improves on previous work for the binary setting (m=2m=2m=2). We also present an inefficient attack that achieves a bias of Ω~(k/n)\tilde\Omega(k/n)Ω~(k/n). Complementing our theoretical work, we give new practical attacks to stress-test multiclass benchmarks by aiming to create as large a bias as possible with a given number of queries. Our experiments show that the additional uncertainty of prediction with a large number of classes indeed mitigates the effect of our best attacks. Our work extends developments in understanding overfitting due to adaptive data analysis to multiclass prediction problems. It also bears out the surprising fact that multiclass prediction problems are significantly more robust to overfitting when reusing a test (or holdout) dataset. This offers an explanation as to why popular multiclass prediction benchmarks, such as ImageNet, may enjoy a longer lifespan than what intuition from literature on binary classification suggests.

View on arXiv
Comments on this paper