Do ImageNet Classifiers Generalize to ImageNet?

Recht, B., Roelofs, R., Schmidt, L., & Shankar, V. (2019). Do ImageNet Classifiers Generalize to ImageNet?. Proceedings of the 36th International Conference on Machine Learning (ICML).

Type

Preprint

Intellectual base

Grey literature

Year

2019

Link

https://proceedings.mlr.press/v97/recht19a

Cited from

statistics_experimental_design

Cited by

2 citations across 2 artifacts.

Each citation links to the sentence it supports in the citing article.

Primes

• Researcher Degrees of Freedom
  • In machine-learning evaluation, tuning on the test set, model selection across many architectures, choice of benchmark suite, and prompt selection each reproduce the structure: many silent comparisons summarised as one.

    This sourceBuilds fresh test sets to probe adaptive overfitting from repeated test-set reuse in ML benchmark evaluation.

Mechanisms

• Baseline Comparison Table
  • The table only measures on the tasks inside it, so a method that wins here can still fail out of distribution — that blind spot belongs to the Challenge Case Red Team and the out-of-distribution monitor, not here.

    This sourceFinds substantial accuracy drops when strong benchmark classifiers are evaluated on newly collected data.

Verification

This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.

Links previously used in the corpus

Before the registry existed this work was also linked 1 other way.

• https://arxiv.org/abs/1902.10811 ×1

Registry ID ref:f28862f37b1a · see in the full table