We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Detecting Overfitting via Adversarial Examples

Abstract: The repeated community-wide reuse of test sets in popular benchmark problems raises doubts about the credibility of reported test-error rates. Verifying whether a learned model is overfitted to a test set is challenging as independent test sets drawn from the same data distribution are usually unavailable, while other test sets may introduce a distribution shift. We propose a new hypothesis test that uses only the original test data to detect overfitting. It utilizes a new unbiased error estimate that is based on adversarial examples generated from the test data and importance weighting. Overfitting is detected if this error estimate is sufficiently different from the original test error rate. We develop a specialized variant of our test for multiclass image classification, and apply it to testing overfitting of recent models to the popular ImageNet benchmark. Our method correctly indicates overfitting of the trained model to the training set, but is not able to detect any overfitting to the test set, in line with other recent work on this topic.
Comments: 17 pages
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Journal reference: Part of: Advances in Neural Information Processing Systems 32 (NIPS 2019) pre-proceedings
Cite as: arXiv:1903.02380 [cs.LG]
  (or arXiv:1903.02380v2 [cs.LG] for this version)

Submission history

From: Roman Werpachowski [view email]
[v1] Wed, 6 Mar 2019 13:49:18 GMT (669kb,D)
[v2] Thu, 14 Nov 2019 11:16:01 GMT (674kb,D)

Link back to: arXiv, form interface, contact.