Evaluating Real-World Generalizability of Algorithm Selection Models ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Computing methodologies Machine learning CCS: Computing methodologies Learning latent representations CCS: Computing methodologies Supervised learning CCS: Theory of computation Design and analysis of algorithms

Gjorgjina Cenikj, Jakub Kudela, Eva Tuba, Tome Eftimov · 2026

Read Evaluating Real-World Generalizability of Algorithm Selection Models ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Computing methodologies Machine learning CCS: Computing methodologies Learning latent representations CCS: Computing methodologies Supervised learning CCS: Theory of computation Design and analysis of algorithms by Gjorgjina Cenikj, Jakub Kudela, Eva Tuba, Tome Eftimov (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-2606-02016.txt.

Description

Cenikj and co-authors test how algorithm selection models generalize to real problems. Algorithm selection picks the best optimizer for an instance. It uses measurable traits and past performance data. Academic suites dominate the training of such models. Real-world landscapes may behave quite differently though.

The study spans two synthetic benchmark suites. BBOB and CEC are the academic sets used. Two real-world problem sets join the evaluation. Robotics trajectory tasks form one real set. Drone path-planning problems form the other. Cross-benchmark tests show where transfer succeeds. They also show where generalization breaks down. Findings inform more reliable selection models for practice. Robustness gaps in current approaches are highlighted.

Catalog record

CREATORGjorgjina Cenikj, Jakub Kudela, Eva Tuba, Tome Eftimov
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.02016

Browse the full library · About the Tower of Babel