Research map / Image generation / Autoregressive and latent generation
BRIC: Bridging Kinematic Plans and Physical Control at Test Time
Dohun Lim, Minji Kim, Jaewoon Lim, Sungchan Kim
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add catastrophic-forgetting loss LCF (ablation) | BRIC · Text-to-motion | success rate: 0.428 -> 0.494 | Improved |
| add robustness loss LRobust (ablation) | BRIC · Text-to-motion | success rate: 0.441 -> 0.494 | Improved |
| add both LCF and LRobust over finetuned CLoSD (ablation) | CLoSD · Text-to-motion | success rate: 0.322 -> 0.494 | Improved |
| test-time adaptation (BRIC) vs CLoSD | CLoSD · Text-to-motion | execution rate: 0.401 -> 0.703 | Improved |
| test-time adaptation (BRIC) vs CLoSD | CLoSD · Text-to-motion | subsequence FID: 19.48 -> 7.96 | Improved |
| add signal-space test-time guidance to baseline without TTA (ablation) | Baseline · Obstacle avoidance (scale 1) | average success rate: 0.80 -> 0.81 | Improved |
| add test-time adaptation to Baseline + TTG (ablation) | Baseline + TTG · Obstacle avoidance (scale 30) | average success rate: 0.43 -> 0.76 | Improved |
| signal-space guidance instead of latent-space guidance (ablation) | BRIC · Obstacle avoidance (scale 30) | average success rate: 0.71 -> 0.76 | Improved |
| signal-space guidance instead of latent-space guidance (ablation) | BRIC · Obstacle avoidance (scale 1) | average success rate: 0.96 vs 0.92 | Mixed |
| signal-space guidance instead of latent-space guidance (ablation) | BRIC · Obstacle avoidance | speed (FPS): 13.58 -> 27.03 | Improved |
Closest papers by meaning
- CLoSD: Closing the Loop between Simulation and Diffusion for multi-task character control (ICLR 2025 · similarity 0.87)
- Efficient Test-Time Adaptation For Robot Policies (NeurIPS 2026 · similarity 0.86)
- UniPhys: Unified Planner and Controller with Diffusion for Flexible Physics-Based Character Control (ICCV 2025 · similarity 0.86)
- PACE: Phase-Aware Chunk Execution for Robot Policies with Action Chunking (NeurIPS 2026 · similarity 0.84)
- Improving Generative Behavior Cloning via Self-Guidance and Adaptive Chunking (NeurIPS 2025 · similarity 0.84)
- One-Step Diffusion Policy: Fast Visuomotor Policies via Diffusion Distillation (ICML 2025 · similarity 0.83)
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.
