English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Learning an accurate model of the environment is essential for model-basedcontrol tasks. Existing methods in robotic visuomotor control usually learnfrom data with heavily labelled actions, object entities or locations, whichcan be demanding in many cases. To cope with this limitation, we propose amethod, dubbed DMotion, that trains a forward model from video data only, viadisentangling the motion of controllable agent to model the transitiondynamics. An object extractor and an interaction learner are trained in anend-to-end manner without supervision. The agent's motions are explicitlyrepresented using spatial transformation matrices containing physical meanings.In the experiments, DMotion achieves superior performance on learning anaccurate forward model in a Grid World environment, as well as a more realisticrobot control environment in simulation. With the accurate learned forwardmodels, we further demonstrate their usage in model predictive control as aneffective approach for robotic manipulations.

DMotion: Robotic Visuomotor Control with Unsupervised Forward Model Learned from Videos