AI Paper Atlas

Research map / Video / Video generation

KineBench: Benchmarking Embodied World Models via IDM-Free Kinematic Grounding

Zeyu Liu, Zhangzhe Zhu, Yang Zhang, Chenyou Fan, Chenjia Bai, Xuelong Li

ECCV 2026 · Video › Video generation · approach: World-model video generation · 5 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 LoRA instead of full-parameter fine-tuningWan2.2 5B · KineBench Suite 0success rate (%): 56.32 -> 19.00Hurt
use LoRA instead of full-parameter fine-tuningWan2.2 5B · KineBench Suite 1success rate (%): 11.90 -> 16.67Improved
use LoRA instead of full-parameter fine-tuningWan2.2 5B · KineBench Suite 2 (unseen assets)success rate (%): 55.50 -> 26.33Hurt
increase training steps from 1.5k to 15k (ablation)Wan2.1 1.3B · KineBench Suite 3success rate (%): 44.83 -> 73.33Improved
increase expert trajectories from scale 10 to scale 50 (ablation)Wan2.1 1.3B · KineBench Suite 3success rate (%): 53.67 -> 51.17No clear effect
increase homogeneous training trajectories from 10 to 100 (ablation)Wan2.1 1.3B · KineBench Suite 3, OpenBoxEasysuccess rate (%): 100% to 83.33%Hurt
increase optimization budget (ablation)Wan2.1 1.3B · KineBench Suite 3, StoreCube-v2success rate (%): 13.34% to 69.52%Improved
evaluate on unseen assets instead of seen assetsWan2.2 · KineBench Suite 2, OpenBoxHardsuccess rate (%): 60.0% to 30.0%Hurt
replace ground-truth depth with MoGeV2-estimated depth (ablation)FoundationPose pipeline · KineBench pipeline pose accuracytranslational error: roughly 1.5--3 cmHurt
use explicit KineBench pipeline instead of IDMaction extraction · Unseen simulator trajectoriesposition error: order of ten centimetersImproved

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