Efficient Swap Regret Minimization in Combinatorial Bandits
Andreas Kontogiannis *, Vasilis Pollatos *, Panayotis Mertikopoulos, Ioannis Panageas · 2026
Read Efficient Swap Regret Minimization in Combinatorial Bandits by Andreas Kontogiannis *, Vasilis Pollatos *, Panayotis Mertikopoulos, Ioannis Panageas (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-2602-02087.txt.
Description
Andreas Kontogiannis and colleagues address fast no-swap regret methods for combinatorial bandits with action count N far beyond dimension. Fast no-swap regret means swap regret growing slower than T with polylog scaling in N. In contrast to weaker external regret, fairly well known, reaching no-swap regret with polylog scaling in N has stayed elusive. Their paper solves this challenge with a no-swap-regret learning method scaling polylog in N, tight for combinatorial bandits. They also show a fast build with per-step cost polylog in N.
Catalog record
| CREATOR | Andreas Kontogiannis *, Vasilis Pollatos *, Panayotis Mertikopoulos, Ioannis Panageas |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 102120 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.02087