AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

K-PWM: Control-Oriented Structured World Models under Partial Observation

Santosh M Rajkumar, Sriram Narayanan, Samuel E Otto, Debdipta Goswami

NeurIPS 2026 · Probabilistic and scientific modeling › Dynamical systems · approach: Data assimilation from sparse observations

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces K-PWM, a Koopman-structured probabilistic world model for control from noisy, partially observed vector measurements of physical systems.

Problem

World models are mostly built from pixel observations, but many physical systems are observed via noisy, partially informative vector measurements, so learning control-ready dynamics under partial observation is harder.

Approach

Uses a Koopman-structured probabilistic latent dynamics model so that the learned world model is structured for control rather than being a generic black-box predictor; details beyond the abstract are not available.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use K-PWM instead of MTS3K-PWM · Hopper/Swimmer/Half Cheetah partial obsprediction RMSE: 0.014 -> 0.0087 (Hopper)Improved
use K-PWM instead of MTS3K-PWM · Half Cheetah partial obsprediction RMSE: 0.14 -> 0.048Improved
use K-PWM instead of MTS3K-PWM · Hopper/Swimmer/Half Cheetahtraining time: 7.6-15.1x speedupImproved
use K-PWM instead of MTS3K-PWM · MuJoCo Menagerie robotstraining time: 1.9-5.6x fasterImproved
use K-PWM with MPPI instead of MTS3 modelK-PWM · MuJoCo Menagerie robotsper-step computation time: 4.3-17.0x speedupImproved
use K-PWM instead of DKK-PWM · MuJoCo Menagerie robotsEE tracking RMSE: Panda 7.3 -> 6.1; ALOHA 2 2.1 -> 2.5Mixed
use K-PWM instead of DKK-PWM · MuJoCo Menagerie robotsper-step computation time: 4.9 -> 5.6 (Panda)Hurt
random generator init instead of Koopman-consistent init (ablation)K-PWM · Swimmer, Half CheetahGECM convergenceHurt
use smaller latent dimension than the selected one (ablation)K-PWM · Half CheetahGECM convergenceHurt
GECM instead of joint SGD training (ablation)K-PWM · MuJoCo predictionprediction accuracyNo clear effect
GECM instead of joint SGD training (ablation)K-PWM · MuJoCo predictiontraining time: several times fasterImproved
bilinear instead of linear latent transition (ablation)K-PWM · MuJoCo predictionprediction RMSEImproved

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