close this message
arXiv smileybones icon

Giving Week!

Show your support for Open Science by donating to arXiv during Giving Week, April 25th-29th.

DONATE
Skip to main content
Cornell University
We gratefully acknowledge support from
the Simons Foundation and member institutions.
arxiv logo > cs > arXiv:2112.04094

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Machine Learning

arXiv:2112.04094 (cs)
[Submitted on 8 Dec 2021]

Title:The Effect of Model Size on Worst-Group Generalization

Authors:Alan Pham, Eunice Chan, Vikranth Srivatsa, Dhruba Ghosh, Yaoqing Yang, Yaodong Yu, Ruiqi Zhong, Joseph E. Gonzalez, Jacob Steinhardt
Download PDF
Abstract: Overparameterization is shown to result in poor test accuracy on rare subgroups under a variety of settings where subgroup information is known. To gain a more complete picture, we consider the case where subgroup information is unknown. We investigate the effect of model size on worst-group generalization under empirical risk minimization (ERM) across a wide range of settings, varying: 1) architectures (ResNet, VGG, or BERT), 2) domains (vision or natural language processing), 3) model size (width or depth), and 4) initialization (with pre-trained or random weights). Our systematic evaluation reveals that increasing model size does not hurt, and may help, worst-group test performance under ERM across all setups. In particular, increasing pre-trained model size consistently improves performance on Waterbirds and MultiNLI. We advise practitioners to use larger pre-trained models when subgroup labels are unknown.
Comments: The first four authors contributed equally to the work
Subjects: Machine Learning (cs.LG)
Cite as: arXiv:2112.04094 [cs.LG]
  (or arXiv:2112.04094v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2112.04094
arXiv-issued DOI via DataCite

Submission history

From: Alan Pham [view email]
[v1] Wed, 8 Dec 2021 03:45:47 UTC (524 KB)
Full-text links:

Download:

  • PDF
  • Other formats
(license)
Current browse context:
cs.LG
< prev   |   next >
new | recent | 2112
Change to browse by:
cs

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar

DBLP - CS Bibliography

listing | bibtex
Yaoqing Yang
Yaodong Yu
Ruiqi Zhong
Joseph E. Gonzalez
Jacob Steinhardt
a export bibtex citation Loading...

Bibtex formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Mendeley logo Reddit logo ScienceWISE logo

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code and Data Associated with this Article

arXiv Links to Code & Data (What is Links to Code & Data?)

Demos

Replicate (What is Replicate?)

Recommenders and Search Tools

Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs and how to get involved.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)
  • About
  • Help
  • contact arXivClick here to contact arXiv Contact
  • subscribe to arXiv mailingsClick here to subscribe Subscribe
  • Copyright
  • Privacy Policy
  • Web Accessibility Assistance
  • arXiv Operational Status
    Get status notifications via email or slack