AI Paper Atlas

Research map / Reinforcement learning / MDPs and planning

Model-based Reinforcement Learning for Semi-Markov Decision Processes with Neural ODEs

Jianzhun Du, Joseph Futoma, Finale Doshi-Velez

NeurIPS 2020 · Reinforcement learning › MDPs and planning · approach: Average-reward and regularized MDPs · 25 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
use Latent-ODE dynamics model instead of vanilla RNNmodel-based RL on HIV · HIVcumulative reward (x10^7): 0.78 -> 30.32Improved
use ODE-RNN instead of vanilla RNNmodel-based RL on windy gridworld · Windy gridworldcumulative reward: -54.02 -> -58.50Hurt
use ODE-RNN instead of vanilla RNNmodel-based RL on Acrobot · Acrobotcumulative reward: -179.35 -> -65.90Improved
lower state prediction error does not imply better policy (Delta t-RNN vs Latent-ODE)Delta t-RNN · Windy gridworldstate prediction error vs reward: 0.334 vs 0.845 error; -45.64 vs -34.87 rewardMixed
transfer pretrained Latent-ODE to new regular time schedule without retrainingLatent-ODE · Acrobot tau=0.2cumulative reward: -54.26 -> -171.37Hurt
add latent variable z to policy (ablation)Latent-ODE model-based policy · Partially observable HIVreward (x10^7): 12.88 -> 18.04Improved
optimize interval times with model-based scheduleHIV Latent-ODE · HIVinteraction rate at similar returnImproved

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