AI Paper Atlas

Research map / 3D vision / 3D generation

gradSim: Differentiable simulation for system identification and visuomotor control

J. Krishna Murthy, Miles Macklin, Florian Golemo, Vikram Voleti, Linda Petrini, Martin Weiss, Breandan Considine, Jérôme Parent-Lévesque, Kevin Xie, Kenny Erleben, Liam Paull, Florian Shkurti and 2 more

ICLR 2021 · 3D vision › 3D generation · approach: Learning physics from video · 43 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
differentiate through physics and rendering to estimate mass from video∇Sim · rigid mass estimationAbs. Rel. err.: 9.01e-5 vs 0.0094Improved
use image-space video supervision instead of precise 3D state supervisionrigid mass estimation · rigid mass estimationAbs. Rel. err.: 9.01e-5 vs 5.17e-9Hurt
estimate contact parameters from video using the unified simulator∇Sim · rigid-body parameter estimationfriction mu Rel. MAE: 0.0107 -> 0.0073Improved
estimate cloth velocity from images rather than 3D state labels∇Sim · cloth parameter estimationper-particle velocity Rel. MAE: 0.127 -> 0.026Improved
omit friction from the dynamics model (ablation)∇Sim · mass estimation with imperfect simulator modelsMean Rel. Abs. Err.: 0.1071 -> 0.1866Hurt
omit elasticity from the dynamics model (ablation)∇Sim · mass estimation with imperfect simulator modelsMean Rel. Abs. Err.: 0.1071 -> 0.2281Hurt
model rigid objects as deformable (ablation)∇Sim · mass estimation with imperfect simulator modelsMean Rel. Abs. Err.: 0.1071 -> 0.3462Hurt
model deformable objects as rigid (ablation)∇Sim · mass estimation with imperfect simulator modelsMean Rel. Abs. Err.: 0.1071 -> 0.4974Hurt
replace differentiable rasterization with photorealistic rendering (ablation)∇Sim · mass estimation with imperfect simulator modelsMean Rel. Abs. Err.: 0.1071 -> 0.1793Hurt
include shading and texture cues instead of using only uniform color (ablation)∇Sim renderer · rigid mass estimationconvergence speedImproved
use texture rather than shaded-only rendering (ablation)∇Sim renderer · rigid mass estimationconvergence speedImproved
use a single visual target image rather than dense 3D supervision∇Sim visuomotor control · control-femgoal attainmentMixed

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