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

An efficient, generalizable physical simulator with universal uncertaintyestimates has wide applications in robot state estimation, planning, andcontrol. In this paper, we build such a simulator for two scenarios, planarpushing and ball bouncing, by augmenting an analytical rigid-body simulatorwith a neural network that learns to model uncertainty as residuals. Combiningsymbolic, deterministic simulators with learnable, stochastic neural netsprovides us with expressiveness, efficiency, and generalizabilitysimultaneously. Our model outperforms both purely analytical and purely learnedsimulators consistently on real, standard benchmarks. Compared with methodsthat model uncertainty using Gaussian processes, our model runs much faster,generalizes better to new object shapes, and is able to characterize thecomplex distribution of object trajectories.

Augmenting Physical Simulators with Stochastic Neural Networks: Case Study of Planar Pushing and Bouncing