Index-equating conjecture for UCB algorithms
Index-equating conjecture for UCB algorithms
Consider a -armed bandit model. For arm , let be the number of pulls by time , let be its empirical mean reward, and let be the index function of a UCB algorithm. In a “reasonable” bandit model and under a “reasonable” UCB algorithm, index-equating conjecture. The pull counts should satisfy
with
This conjecture informally extends the equal-index characterization of the associated fluid system to the stochastic bandit process. The paper presents it as a natural conjecture within a perturbation-analysis framework; no resolution is supplied in the provided text.
Sources & referencesView supporting material
Primary source
Yilun Chen and Jiaqi Lu, “A characterization of sample adaptivity in UCB data”, arXiv:2503.04855 (2025).
Progress summary
Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.
Solutions 0
Sign in to submit a solution.
No solutions have been posted yet.