Cryo-Bench: Benchmarking Foundation Models for Cryosphere Applications
Saurabh Kaushik, Lalit Maurya, Beth Tellman · 2026
Read Cryo-Bench: Benchmarking Foundation Models for Cryosphere Applications by Saurabh Kaushik, Lalit Maurya, Beth Tellman (2026) 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-2603-01576.txt.
Description
Earth models that map the planet from space have had little fair testing on ice and snow regions. Saurabh Kaushik, Lalit Maurya, and Beth Tellman fill that gap with a new test set in this 2026 work. Their Cryo-Bench covers glaciers with debris, glacial lakes, sea ice, and glacier fronts across many regions and sensors.
The team scores fourteen large Earth models against two standard image baselines. With frozen cores, a classic baseline still leads on mean overlap score, while top Earth models sit close behind. With just a tenth of the training data, Earth models pull ahead, which suits sparse polar labels. Full tuning gives mixed gains unless learning rates are tuned too, a practical tip the authors stress for users.
Catalog record
| CREATOR | Saurabh Kaushik, Lalit Maurya, Beth Tellman |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43365 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.01576