Dimensions overall score 2.0

GitHub Code Pulse

No public code linked for this paper yet.

• The paper adapts the policy gradient analysis for continuous time k-armed stochastic bandits to a discrete time setup.method 95%
• With a specific learning rate, the regret for stochastic bandits is bounded.result 95%
• The regret bound is O(k log(k) log(n) / η) for a horizon n and learning rate η.result 95%
• The learning rate required for the regret bound is η = O(Δ_min^2 / (Δ_max log(n))).technical 95%
• The analysis is based on a Lyapunov approach.method 95%
• The problem setting involves k actions and a horizon n, where n is greater than or equal to k.technical 95%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card