Distributionally Robust Coreset Selection under Covariate Shift
Tomonari Tanaka, Hanting Yang, Tatsuya Aoyama, Satoshi Akahane, Yoshito Okura Nagoya University, Hiroyuki Hanada, Noriaki Hashimoto RIKEN, Yu Inatsu Nagoya Institute of Technology, Taro Murayama, Hanju Lee, Shinya Kojima DENSO CORPORATION, Ichiro Takeuchi Nagoya University & RIKEN takeuchi.ichiro.n6 · 2025
Read Distributionally Robust Coreset Selection under Covariate Shift by Tomonari Tanaka, Hanting Yang, Tatsuya Aoyama, Satoshi Akahane, Yoshito Okura Nagoya University, Hiroyuki Hanada, Noriaki Hashimoto RIKEN, Yu Inatsu Nagoya Institute of Technology, Taro Murayama, Hanju Lee, Shinya Kojima DENSO CORPORATION, Ichiro Takeuchi Nagoya University & RIKEN takeuchi.ichiro.n6 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-14253.txt.
Description
Tomonari Tanaka and co-authors present a robust way to shrink training data in this 2025 paper. Coreset selection picks a small subset of training examples that still teaches a model well. The catch is shift: data at deployment rarely matches data at training time, and the future spread is unknown.
Their method, called DRCS, plans for the worst shift inside a set range. It derives a bound on the worst-case test error under that shift. The coreset is then chosen to keep that bound low, rather than to fit today's data alone. The approach suits fields where conditions drift between lab and field use. A subset chosen this way should hold up across many futures, not just the one the training set happens to show.
Catalog record
| CREATOR | Tomonari Tanaka, Hanting Yang, Tatsuya Aoyama, Satoshi Akahane, Yoshito Okura Nagoya University, Hiroyuki Hanada, Noriaki Hashimoto RIKEN, Yu Inatsu Nagoya Institute of Technology, Taro Murayama, Hanju Lee, Shinya Kojima DENSO CORPORATION, Ichiro Takeuchi Nagoya University & RIKEN takeuchi.ichiro.n6 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-14253