SpectralTS achieves regret scaling as d sqrt(T ln N) for graph-smooth bandit problems, matching known bounds while scaling better for large numbers of choices.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Spectral Thompson sampling
SpectralTS achieves regret scaling as d sqrt(T ln N) for graph-smooth bandit problems, matching known bounds while scaling better for large numbers of choices.