Research map / Reinforcement learning / Visual and representation RL
FoAM: Foresight-Augmented Multi-Task Imitation Policy for Robotic Manipulation
Litao Liu, Wentao Wang, Yifan Han, Zhuoli Xie, Pengfei Yi, Junyan Li, Wenzhao Lian
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 foresight augmentation (foresight loss) (ablation) | FoAM · FoAM benchmark Dual-Arm | success rate: 36% -> 86% | Improved |
| add foresight augmentation (ablation) | FoAM · FoAM benchmark Block-based | success rate: 81% -> 91% | Improved |
| add foresight augmentation (ablation) | FoAM · FoAM benchmark Unseen Tasks | success rate: 11% -> 66% | Improved |
| add foresight augmentation (ablation) | FoAM · Real-world scenarios | successes: 11/40 vs 7/40 in Scenario I | Improved |
| add foresight augmentation (ablation) | FoAM · Real-world Scenario II | successes: 13/40 -> 11/40 | Hurt |
| add task instruction to goal-image policy (multi-modal condition) without foresight (ablation) | Gimg-ACT · FoAM benchmark Unseen Tasks | success rate: 45% -> 11% | Hurt |
| add task instruction to goal-image policy without foresight (ablation) | Gimg-ACT · FoAM benchmark Block-based | success rate: 39% -> 81% | Improved |
| use goal images instead of language only | MT-ACT/BAKU · FoAM benchmark Unseen Tasks | success rate: 0 -> 45% (Gimg-ACT) | Improved |
| use multi-modal goal condition (language + goal image) | FoAM · FoAM benchmark Dual-Arm | success rate: 45% -> 86% | Improved |
| replace real goal images with VLM-generated goal images (ablation) | FoAM · Real-world Scenario III Bitter Melon | successes: 5/10 -> 7/10 | Improved |
| replace real goal images with VLM-generated goal images (ablation) | FoAM · Real-world Scenario III Eggplant | successes: 7/10 -> 7/10 | No clear effect |
| replace real goal images with VLM-generated goal images (ablation) | FoAM · Real-world Scenario III Peach/Tomato | successes: 2/10 -> 3/10 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- BAKU — BAKU: An Efficient Transformer for Multi-Task Policy Learning (NeurIPS 2024) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- AffordDP: Generalizable Diffusion Policy with Transferable Affordance (CVPR 2025 · similarity 0.88)
- What's the Move? Hybrid Imitation Learning via Salient Points (ICLR 2025 · similarity 0.87)
- Multi-Stage Manipulation with Demonstration-Augmented Reward, Policy, and World Model Learning (ICML 2025 · similarity 0.87)
- Dream to Manipulate: Compositional World Models Empowering Robot Imitation Learning with Imagination (ICLR 2025 · similarity 0.87)
- GravMAD: Grounded Spatial Value Maps Guided Action Diffusion for Generalized 3D Manipulation (ICLR 2025 · similarity 0.87)
- VIP: Vision Instructed Pre-training for Robotic Manipulation (ICML 2025 · similarity 0.87)
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.
