AI Paper Atlas

Research map / Video / Video generation

InsActor: Instruction-driven Physics-based Characters

Jiawei Ren, Mingyuan Zhang, Cunjun Yu, Xiao Ma, Liang Pan, Ziwei Liu

NeurIPS 2023 · Video › Video generation · approach: Text-to-motion diffusion · 3 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 hierarchical diffusion planning with low-level skill mapping instead of DReCon trackingphysics-based instruction-driven animation · KIT-ML testR Precision Top 3: 0.470 -> 0.648Improved
use hierarchical diffusion planning with low-level skill mapping instead of DReCon trackingphysics-based instruction-driven animation · HumanML3D testFID: 1.244 -> 0.566Improved
retain the hierarchical method under physical perturbationsInsActor · KIT-ML test with perturbationR Precision Top 3: 0.447 -> 0.599Improved
apply waypoint inpainting while retaining the language conditionInsActor · HumanML3D waypoint headingSuccess rate: 0.380 -> 0.907Improved
remove the language condition (ablation)InsActor waypoint heading · HumanML3DR Precision Top 3: 0.388 -> 0.089Hurt
remove the waypoint condition (ablation)InsActor waypoint heading · HumanML3DSuccess rate: 0.907 -> 0.081Hurt
add the low-level skill policy to high-level diffusion planning (ablation)InsActor · KIT-ML testR Precision Top 3: 0.188 -> 0.648Improved
add the low-level skill policy to high-level diffusion planning (ablation)InsActor · KIT-ML testFID: 1.106 -> 0.786Improved
add high-level diffusion planning to the low-level policy (ablation)InsActor · KIT-ML testR Precision Top 3: 0.460 -> 0.648Improved
increase the skill-space compactness weight (ablation)InsActor · KIT-MLinstruction-motion alignmentHurt
decrease the skill-space compactness weight (ablation)InsActor · KIT-MLmotion plausibilityHurt
smooth the planned motions (ablation)DReCon tracking of generated plans · HumanML3D testFID: 1.244 -> 1.271Hurt

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