README.md

---
license: cc-by-4.0
language:
- en
pretty_name: WOVEN
task_categories:
- visual-question-answering
- image-text-to-text
tags:
- world-model
- visual-reasoning
- physical-reasoning
- spatial-reasoning
- temporal-reasoning
- video-generation
- synthetic
size_categories:
- 10K<n<100K
viewer: false
---

<h1 align="center">WOVEN: Weaving Visual World Modeling into Multimodal LLMs</h1>

<p align="center">
  <a href="https://arxiv.org/abs/2610.12417"><img src="https://img.shields.io/badge/📄_Paper-arXiv_2610.12417-B31B1B?style=for-the-badge" alt="Paper"></a>
  <a href="https://woven-ai.github.io"><img src="https://img.shields.io/badge/🏠_Project_Page-2453C4?style=for-the-badge" alt="Project page"></a>
  <a href="https://github.com/mll-lab-nu/WOVEN"><img src="https://img.shields.io/badge/💻_Code-181717?style=for-the-badge" alt="Code"></a>
</p>

> [!NOTE]
> **October 9, 2026:** The training split is available now. We are uploading the validation split and the three test sets, and they will be available here soon.

WOVEN is a training source and benchmark for **visual transition reasoning**. A transition (s, a, s′) consists of the visual state of a scene before an action, the action, and the state after it. Visual transition reasoning infers the unobserved part of a transition from its observed parts: predicting an outcome, identifying the action behind a change, reasoning about an alternative action, or ordering intermediate states.

WOVEN organizes transition supervision by scene, action, and reasoning type, using diverse, realistic rollouts from a video generation model: **36,076** four-option multiple-choice items across **20 scene types**, **5 action types**, and **8 reasoning types**. In the paper, training subsets of only about 2,000 WOVEN items each collectively improve 22 of 26 external benchmarks, by up to 27.3 percentage points.

<p align="center">
  <img src="https://huggingface.co/datasets/MLL-Lab/WOVEN/resolve/main/assets/teaser.jpg" width="900" alt="Overview. (1) MLLMs fail across spatial, embodied, physical, and temporal tasks. (2) These failures share a weakness: reasoning about how visual states, actions, and outcomes relate. (3) WOVEN generates (s, a, s′) transitions with a video generation model, organizes them along controlled axes of reasoning, action, and scene, and turns them into error-typed multiple-choice supervision. (4) Training on WOVEN improves 22 of 26 external benchmarks.">
</p>

## Splits

| Split | Items | Status |
|---|---:|---|
| Training | 22,728 | Available |
| Validation | 2,508 | Uploading soon |
| In-distribution test | 6,228 | Uploading soon |
| Held-out-scene test (restaurant and beach scenes) | 3,496 | Uploading soon |
| State-perturbation test | 1,116 | Uploading soon |
| **Total** | **36,076** | |

## Quick start

```python
from datasets import load_dataset

ds = load_dataset("MLL-Lab/WOVEN", data_files="train/questions-*.parquet", split="train")

ex = ds[0]
print(ex["reasoning_type"], ex["action_type"], ex["scene"])
print(ex["question"])                          # images appear as <image_1>, <image_2>, ...
for opt in ex["options"]:
    print(f'{opt["label"]}. {opt["content"]}')  # image options are <image_k> as well
print("answer:", ex["answer"])
ex["images"][0]                                # PIL.Image
```

Turn an item into chat messages with the images in reading order:

```python
import re

def to_messages(ex, instruction="Answer with the option's letter."):
    # The instruction is an example; the prompt templates used in the paper will be released with the code.
    text = ex["question"] + "\n" + "\n".join(f'{o["label"]}. {o["content"]}' for o in ex["options"])
    text += "\n" + instruction
    content = []
    for i, piece in enumerate(re.split(r"<image_(\d+)>", text)):
        if i % 2:
            content.append({"type": "image", "image": ex["images"][int(piece) - 1]})
        elif piece:
            content.append({"type": "text", "text": piece})
    return [{"role": "user", "content": content}]
```

Select items by their labels without decoding any images:

```python
fwd = ds.filter(lambda t: t == "forward_dynamics", input_columns="reasoning_type")
```

Each training rollout is also available as 11 frames sampled every 0.5 seconds. Items and rollouts share `source_id` and `action_id`:

```python
rollouts = load_dataset("MLL-Lab/WOVEN", data_files="rollout_frames/*.parquet", split="train")
r = rollouts[0]
print(r["source_id"], r["action_id"], r["action"])
r["frames"]                                    # 11 PIL images at r["timestamps"] = 0.0, 0.5, ..., 5.0 s
```

To download all files:

```bash
hf download MLL-Lab/WOVEN --repo-type dataset --local-dir WOVEN
```

## Repository layout

```
train/questions-0000{0..4}-of-00005.parquet          22,728 training items, images embedded (2.5 GB)
rollout_frames/rollouts-0000{0..2}-of-00003.parquet  5,172 training rollouts, 11 frames each (1.5 GB)
assets/                                              figures used on this page
```

## Data fields

### Items (`train/`)

| Field | Type | Description |
|---|---|---|
| `id` | string | Item identifier |
| `source_id` | string | Initial state; all rollouts from the same initial frame share it |
| `action_id` | string | The rollout the item is built from (`action_0`, `action_1`, ...) |
| `action_type` | string | `exogenous` (passive physical event), `perceptive` (camera motion), `inspective` (object inspection), `navigative` (navigation), or `manipulative` (object manipulation) |
| `reasoning_type` | string | One of the eight reasoning types below |
| `scene` | string | One of the 20 scene types |
| `scene_group` | string | `indoors` or `outdoors` |
| `physical_principle` | string | For passive physical events: `permanence`, `cohesion`, `solidity`, `gravity`, `support`, `inertia`, `collision`, or `containment`; otherwise null |
| `manipulation_subject` | string | For manipulation: `human` or `humanoid`; otherwise null |
| `manipulation_view` | string | For manipulation: `egocentric` or `allocentric`; otherwise null |
| `question` | string | Question text; images are referenced as `<image_k>` |
| `options` | list | Four options, each with `label` (A to D), `content` (text, or `<image_k>` for an image option), `is_image`, `is_correct`, `distractor_type` (the error type of a wrong option; `correct` for the answer), `action_id` (the rollout an option comes from; `_static_` for the no-change option), `order` (the state order, for temporal ordering), and `is_static` (true for the no-change option) |
| `answer` | string | Letter of the correct option |
| `images` | list of images | All images the item shows, in order of first appearance (question, then options A to D); JPEG, 512 pixels wide |

### Rollouts (`rollout_frames/`)

| Field | Type | Description |
|---|---|---|
| `source_id`, `action_id` | string | Keys shared with the items |
| `action` | string | Description of the action that produced the rollout |
| `timestamps` | list of floats | 0.0, 0.5, ..., 5.0 seconds |
| `frames` | list of images | The 11 frames; JPEG, 512 pixels wide |

## Taxonomy

<p align="center">
  <img src="https://huggingface.co/datasets/MLL-Lab/WOVEN/resolve/main/assets/taxonomy.jpg" width="900" alt="The WOVEN taxonomy. Left: the passive physical event type with its eight principles and the four agent-driven action types. Middle: the eight reasoning types in four families, each marking which element of (s, a, s′) the question asks for. Right: example scenes; restaurant and beach are held out.">
</p>

### Reasoning types

| Family | Reasoning type | `reasoning_type` | The question asks | Training items |
|---|---|---|---|---:|
| Causal dynamics | Forward dynamics | `forward_dynamics` | What will the scene look like after the action? | 4,132 |
| Causal dynamics | Inverse dynamics | `inverse_dynamics` | Which action caused the change? | 4,132 |
| Counterfactual reasoning | Counterfactual removal | `counterfactual_removal` | What would the scene look like if the action had not happened? | 2,060 |
| Counterfactual reasoning | Counterfactual substitution | `counterfactual_substitution` | What would the outcome be under a different action? | 2,060 |
| Physical modeling | Outcome prediction | `outcome_prediction` | How will a passive physical event end? | 1,040 |
| Physical modeling | Cued prediction | `cued_prediction` | How will the event end, given a cue naming the physical principle? | 1,040 |
| Temporal coherence | Temporal ordering | `temporal_ordering` | In what order did these states occur? | 4,132 |
| Temporal coherence | Temporal adjacency | `temporal_adjacency` | Which states come right before and after a reference state? | 4,132 |

### Action types

| `action_type` | Action type | Training items | Additional labels |
|---|---|---:|---|
| `exogenous` | Passive physical events | 4,160 | `physical_principle`: permanence, cohesion, solidity (object properties); gravity, support, inertia, collision, containment (physical events) |
| `perceptive` | Camera motion | 4,144 | |
| `inspective` | Object inspection | 6,120 | |
| `navigative` | Navigation | 4,144 | |
| `manipulative` | Object manipulation | 4,160 | `manipulation_subject` (human, humanoid) and `manipulation_view` (egocentric, allocentric) |

### Scenes

- **Indoors:** bathroom, bedroom, classroom, gym, hospital, kitchen, living room, office, restaurant, supermarket
- **Outdoors:** beach, construction site, crossroad, farm, garage, park, playground, school campus, sidewalk, yard

Restaurant and beach are held out: they appear only in the held-out-scene test.

## How the data was built

- Rollouts are generated with a video generation model (Wan2.2-I2V-A14B) from specified initial frames and action descriptions. The resulting and intermediate states come from the rollout rather than from the conditioning prompt.
- Several rollouts under different actions start from the same initial state, so alternative outcomes of the same scene are available.
- Each rollout lasts 5 seconds, and 11 frames are sampled every 0.5 seconds; the questions are built from these frames.
- Each question is phrased with one of 30 templates drawn at random.
- The three wrong options are usually outcomes of other rollouts from the same initial state, and each is labeled with its error type. For physical-modeling items, the wrong options are produced by an image editor (Gemini 3.1 Flash Image) from a closed inventory of physically violating edit types.
- Images are stored as 512-pixel-wide JPEGs, the resolution used for training in the paper.

See the [paper](https://arxiv.org/abs/2610.12417) for the full construction and quality-control procedure.

## Intended use

The training split is intended for training and analyzing visual transition reasoning in multimodal models. The test sets, once released, are intended for evaluation only; please do not train on them.

## License

The dataset is released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).

## Citation

```bibtex
@article{fan2026woven,
  title={WOVEN: Weaving Visual World Modeling into Multimodal LLMs},
  author={Fan, Zheyu and Zhang, Yue and Deng, Mingkai and Wang, Kangrui and Wang, Qineng and Chen, Canyu and Hao, Jie and Fan, Xing and Guo, Chenlei and Xing, Eric P. and Bansal, Mohit and Li, Manling},
  journal={arXiv preprint arXiv:2610.12417},
  year={2026}
}
```

For questions, please open a discussion on this page.