Hedging on the Frontier: Learning New Tasks with Few Samples

Tobias Wegel, Federico Di Gennaro, Geelon So ∗, Fanny Yang · 2026

Read Hedging on the Frontier: Learning New Tasks with Few Samples by Tobias Wegel, Federico Di Gennaro, Geelon So ∗, Fanny Yang (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-2605-30997.txt.

Description

Wegel and colleagues learn new tasks from few samples by exploiting benchmark scores on related tasks. They observe weak monotonicity: a model dominating rivals on many benchmarks tends to win on the new task too. Under approximate weak monotonicity, they prune the model class and adapt to the trade-off geometry by hedging on the frontier, within both transfer learning and model-selection aggregation. The framework turns benchmark evaluations into provable gains for few-sample learning.

Catalog record

CREATORTobias Wegel, Federico Di Gennaro, Geelon So ∗, Fanny Yang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE275838
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30997

Browse the full library · About the Tower of Babel