Add WAM paper: DSWAM: A Dual-System World Action Foundation Model for Fine-Grained Robot Manipulation

[279object]

Paper

• Title: DSWAM: A Dual-System World Action Foundation Model for Fine-Grained Robot Manipulation
• Short name: DSWAM
• arXiv ID: 2607.04927v1
• Paper: https://arxiv.org/pdf/2607.04927
• Authors: Jian Zhu, Jianjun Zhang, Taiyi Su, Tianbin Liu, Zhangyuan Wang, Kai Xie, Zitai Huang, Chong Ma, Youzhang He, Tianjian Wang, Hanyang Wang, Weihao Ding, Yi Xu
• Published: 2026-07-06
• arXiv categories: cs.RO, cs.AI
• Matched keywords: WAM, world action model, world action models

README Entry

- **DSWAM**: "DSWAM: A Dual-System World Action Foundation Model for Fine-Grained Robot Manipulation", arXiv 2026. ![](https://img.shields.io/badge/Unified--DiT-be123c) ![](https://img.shields.io/badge/Explicit-f43f5e)
  [[📄 Paper](https://arxiv.org/pdf/2607.04927)]

Robot Decision

• Decision: accept
• Confidence: high
• Taxonomy: joint_diffusion_unified_explicit
• Taxonomy confidence: medium
• README heading: ### Diffusion-based Generation
• Badges: https://img.shields.io/badge/Unified--DiT-be123c, https://img.shields.io/badge/Explicit-f43f5e

Reason

The paper explicitly introduces and evaluates a World Action Model framework for robot manipulation, with action-conditioned world-aware execution and real-robot control.

Evidence

• DSWAM: Dual-System World Action Foundation Model for fine-grained robot manipulation
• WAM executor trained with action prediction and video co-training for robot action chunks
• evaluated on real-world deformable manipulation and RoboTwin tasks

Taxonomy Evidence

• executor is trained with action prediction and video co-training
• inference directly predicts action chunks without explicit future video generation

Human Review Checklist

• This paper belongs in the WAM survey
• README section is correct
• Badges are correct
• Short name is correct
• Paper link is correct
• Merge this PR if accepted