AI Paper Atlas

Research map / Bandits and online learning / RL theory and MDPs

Maximum Expected Hitting Cost of a Markov Decision Process and Informativeness of Rewards

Falcon Dai, Matthew Walter

NeurIPS 2019 · Bandits and online learning › RL theory and MDPs · approach: Average-reward and latent MDPs · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Replace the diameter-based value-span bound with the MEHC boundUCRL2-style regret analysis · MDPs whose actual model lies in the extended MDPoptimal value span upper bound: r_\text{max} D(M) -> \kappa(M)Improved
Use MEHC instead of diameter in the regret boundUCRL2 · Average-reward MDPsregret upper bound: 34 \max\{1, \kappa\} S \sqrt{AT \log\left(\frac{T}{\delta}\right)}Improved
Apply potential-based reward shaping to make transition-relevant actions more distinguishable in the toy MDPTwo-state toy MDP · Two-state toy MDPMEHC: \frac{\alpha + \beta}{2 \epsilon} < \frac{\alpha}{\epsilon}Improved
Apply any bounded potential-based reward shaping transformationMDP with finite MEHC and unsaturated optimal average reward · MDPs with finite MEHC and unsaturated optimal average rewardMEHC: \frac{1}{2} \kappa(M) \leq \kappa(M^\varphi) \leq 2 \kappa(M)Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy