# Do-Undo Bench: Reversibility for Action Understanding in Image Generation

arXiv: https://arxiv.org/abs/2512.13609 · HTML: https://mlsanity.com/paper/2512.13609/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Embodied and 3D vision-language models, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 28 of its 33 references that are in the map.
- First posted on arXiv 15 Dec 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 28 of 28)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow](https://mlsanity.com/paper/2209.03003/) (2022)
- [InstructPix2Pix: Learning to Follow Image Editing Instructions](https://mlsanity.com/paper/2211.09800/) (2022)
- [RAFT: Recurrent All-Pairs Field Transforms for Optical Flow](https://mlsanity.com/paper/2003.12039/) (2020)
- [FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space](https://mlsanity.com/paper/2506.15742/) (2025)
- [Emerging Properties in Unified Multimodal Pretraining](https://mlsanity.com/paper/2505.14683/) (2025)
- [DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset](https://mlsanity.com/paper/2403.12945/) (2024)
- [MagicBrush: A Manually Annotated Dataset for Instruction-Guided Image Editing](https://mlsanity.com/paper/2306.10012/) (2023)
- [BLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and Dataset](https://mlsanity.com/paper/2505.09568/) (2025)
- [Emu Edit: Precise Image Editing via Recognition and Generation Tasks](https://mlsanity.com/paper/2311.10089/) (2023)
- [ImgEdit: A Unified Image Editing Dataset and Benchmark](https://mlsanity.com/paper/2505.20275/) (2025)
- [The EPIC-KITCHENS Dataset: Collection, Challenges and Baselines](https://mlsanity.com/paper/2005.00343/) (2020)
- [How Far is Video Generation from World Model: A Physical Law Perspective](https://mlsanity.com/paper/2411.02385/) (2024)
- [Learning Video Representations from Large Language Models](https://mlsanity.com/paper/2212.04501/) (2022)
- [HQ-Edit: A High-Quality Dataset for Instruction-based Image Editing](https://mlsanity.com/paper/2404.09990/) (2024)
- [Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation](https://mlsanity.com/paper/2410.05363/) (2024)
- [Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning](https://mlsanity.com/paper/2503.15558/) (2025)
- [PhyScene: Physically Interactable 3D Scene Synthesis for Embodied AI](https://mlsanity.com/paper/2404.09465/) (2024)
- [GoT: Unleashing Reasoning Capability of Multimodal Large Language Model for Visual Generation and Editing](https://mlsanity.com/paper/2503.10639/) (2025)
- [Synthetic Experience Replay](https://mlsanity.com/paper/2303.06614/) (2023)
- [SEED-Data-Edit Technical Report: A Hybrid Dataset for Instructional Image Editing](https://mlsanity.com/paper/2405.04007/) (2024)
- [GenHowTo: Learning to Generate Actions and State Transformations from Instructional Videos](https://mlsanity.com/paper/2312.07322/) (2023)
- [Learning Action and Reasoning-Centric Image Editing from Videos and Simulations](https://mlsanity.com/paper/2407.03471/) (2024)
- [UniBench: Visual Reasoning Requires Rethinking Vision-Language Beyond Scaling](https://mlsanity.com/paper/2408.04810/) (2024)
- [PICABench: How Far Are We from Physically Realistic Image Editing?](https://mlsanity.com/paper/2510.17681/) (2025)
- [Visual Jenga: Discovering Object Dependencies via Counterfactual Inpainting](https://mlsanity.com/paper/2503.21770/) (2025)
- [Action-based image editing guided by human instructions](https://mlsanity.com/paper/2412.04558/) (2024)

## Abstract (arXiv)

We introduce the Do-Undo task and benchmark to address a critical gap in vision-language models: understanding and generating plausible scene transformations driven by real-world actions. Unlike prior work that relies on prompt-based image generation and editing to perform action-conditioned image manipulation, our training hypothesis requires models to simulate the outcome of a real-world action and then reverse it to the original state. This forward-reverse requirement tests genuine cause-and-effect understanding rather than stylistic or semantic edits. We curate a high-quality benchmark of reversible actions from real-world scenarios to enable robust action grounding. Our experiments reveal that current models struggle with action reversibility, highlighting the need to evaluate action understanding. Do-Undo provides an intuitive testbed for evaluating and advancing action-aware generation in multimodal systems that must reason about real-world dynamics.