Generalization Error
Generalization Error is the expected value of the discrepancy between a model's predicted outputs and the true outputs on unseen data, serving as a measure of the model's ability to perform accurately on new instances drawn from the same underlying distribution, thereby distinguishing it from training error and reflecting the effectiveness of learning in the context of supervised learning, overfitting, underfitting, model selection, and the bias-variance tradeoff.
1.2K
Publications
98.6K
Citations
2.9K
Authors
867
Institutions
Publications per year
2017–2026
597
Variants
Generalisation Error
1.2K
Understanding deep learning (still) requires rethinking generalization
Chiyuan Zhang, Samy Bengio, Moritz Hardt et al. · Communications of the ACM · 2021 · 2.3K citations · Full text
1–5 of 1.2K
2.9K
19
h-index
16
h-index
15
h-index
12
h-index
12
h-index
1–5 of 2.9K
867
University of California, Berkeley
Berkeley, United States
22 h-index
49 publications
11.9K citations
Google (United States)
Mountain View, United States
22 h-index
67 publications
11.7K citations
Stanford University
Stanford, United States
21 h-index
61 publications
5.3K citations
Tokyo Institute of Technology
Tokyo, Japan
18 h-index
37 publications
1.8K citations
Carnegie Mellon University
Pittsburgh, United States
18 h-index
44 publications
1.6K citations

The University of Sydney
Sydney, Australia
18 h-index
39 publications
2K citations
The University of Tokyo
Tokyo, Japan
17 h-index
43 publications
4.1K citations
University of Illinois Urbana-Champaign
Urbana, United States
16 h-index
39 publications
1.4K citations
1–8 of 867
Repository · 265 publications · 15.5K citations
63
h-index
40 publications · 6.5K citations
30
h-index
TPAMI · 23 publications · 6.5K citations
22
h-index
Conference · 24 publications · 1.6K citations
18
h-index
Conference · 24 publications · 981 citations
16
h-index
1–5