Scientific Argument with Supervised Learning

Abstract

The use of machine learning (ML) for scientific discovery has enabled data-driven approaches to new and old questions alike. We argue that scientific arguments based on algorithms for discovery hold the potential to reinforce existing assumptions about phenomena, under the guise of testing them. Using examples from image-based biological classification, we show how scientific arguments using supervised learning can contribute to unintended, unrealistic, or under-evidenced claims.

PDF NeurIPSW OpenReview Semantic Scholar

Cite

Text

Lockhart and Jacobs. "Scientific Argument with Supervised Learning." NeurIPS 2021 Workshops: AI4Science, 2021.

Markdown

[Lockhart and Jacobs. "Scientific Argument with Supervised Learning." NeurIPS 2021 Workshops: AI4Science, 2021.](https://mlanthology.org/neuripsw/2021/lockhart2021neuripsw-scientific/)

BibTeX

@inproceedings{lockhart2021neuripsw-scientific,
  title     = {{Scientific Argument with Supervised Learning}},
  author    = {Lockhart, Jeffrey W and Jacobs, Abigail Z.},
  booktitle = {NeurIPS 2021 Workshops: AI4Science},
  year      = {2021},
  url       = {https://mlanthology.org/neuripsw/2021/lockhart2021neuripsw-scientific/}
}