README.md

---
license: cc-by-4.0
language:
- en
size_categories:
- 100K<n<1M
pretty_name: Causal3D
tags:
- Causality
- Computer_Vision
dataset_info:
- config_name: hypothetical_scenes_Hypothetic_v2_linear
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2197142
    num_examples: 14368
  download_size: 0
  dataset_size: 2197142
- config_name: hypothetical_scenes_Hypothetic_v2_nonlinear
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 1809956
    num_examples: 10000
  download_size: 0
  dataset_size: 1809956
- config_name: hypothetical_scenes_Hypothetic_v3_fully_connected_linear
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 1397093
    num_examples: 10000
  download_size: 0
  dataset_size: 1397093
- config_name: hypothetical_scenes_Hypothetic_v4_linear_full_connected
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 1699598
    num_examples: 10050
  download_size: 0
  dataset_size: 1699598
- config_name: hypothetical_scenes_Hypothetic_v4_linear_v
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2053379
    num_examples: 10000
  download_size: 0
  dataset_size: 2053379
- config_name: hypothetical_scenes_Hypothetic_v4_nonlinear_v
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2828217
    num_examples: 10000
  download_size: 0
  dataset_size: 2828217
- config_name: hypothetical_scenes_Hypothetic_v5_linear
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 1956461
    num_examples: 10000
  download_size: 0
  dataset_size: 1956461
- config_name: hypothetical_scenes_Hypothetic_v5_linear_full_connected
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 1955921
    num_examples: 10000
  download_size: 0
  dataset_size: 1955921
- config_name: hypothetical_scenes_rendered_h3_linear_128P
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 5425498
    num_examples: 15000
  download_size: 0
  dataset_size: 5425498
- config_name: hypothetical_scenes_rendered_h3_nonlinear_128P
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 3239120
    num_examples: 10223
  download_size: 0
  dataset_size: 3239120
- config_name: hypothetical_scenes_rendered_h5_nonlinear
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 5459126
    num_examples: 10360
  download_size: 0
  dataset_size: 5459126
- config_name: real_scenes_Real_Parabola
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 1323548
    num_examples: 10000
  download_size: 0
  dataset_size: 1323548
- config_name: real_scenes_Real_magnet_v3
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 28397
    num_examples: 481
  download_size: 0
  dataset_size: 28397
- config_name: real_scenes_Real_magnet_v3_5
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 93977
    num_examples: 1503
  download_size: 0
  dataset_size: 93977
- config_name: real_scenes_Real_parabola_multi_view
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 0
    num_examples: 0
  download_size: 0
  dataset_size: 0
- config_name: real_scenes_Real_spring_v3_256P
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 136325
    num_examples: 450
  download_size: 0
  dataset_size: 136325
- config_name: real_scenes_Water_flow_scene_render
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2792618
    num_examples: 10000
  download_size: 0
  dataset_size: 2792618
- config_name: real_scenes_convex_len_render_images
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 72448
    num_examples: 1078
  download_size: 0
  dataset_size: 72448
- config_name: real_scenes_real_pendulum
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2925963
    num_examples: 9999
  download_size: 0
  dataset_size: 2925963
- config_name: real_scenes_rendered_magnetic_128
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2324526
    num_examples: 8350
  download_size: 0
  dataset_size: 2324526
- config_name: real_scenes_rendered_reflection_128P
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2765222
    num_examples: 9995
  download_size: 0
  dataset_size: 2765222
- config_name: real_scenes_seesaw_scene_128P
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2275814
    num_examples: 10000
  download_size: 0
  dataset_size: 2275814
- config_name: real_scenes_spring_scene_128P
  features:
  - name: image
    dtype: image
  - name: file_name
    dtype: string
  - name: metadata
    dtype: string
  splits:
  - name: train
    num_bytes: 2547386
    num_examples: 10000
  download_size: 0
  dataset_size: 2547386
---
# 🧠 Causal3D: A Benchmark for Visual Causal Reasoning

**Causal3D** is a comprehensive benchmark designed to evaluate models’ abilities to uncover *latent causal relations* from structured and visual data. This dataset integrates **3D-rendered scenes** with **tabular causal annotations**, providing a unified testbed for advancing *causal discovery*, *causal representation learning*, and *causal reasoning* with **vision-language models (VLMs)** and **large language models (LLMs)**.

## 📚 Usage

#### 🔹 Option 1: Load from Hugging Face

You can easily load a specific scene using the Hugging Face `datasets` library:

```python
from datasets import load_dataset

dataset = load_dataset(
    "LLDDSS/Causal3D",
    name="real_scenes_Real_Parabola",
    download_mode="force_redownload",  # Optional: force re-download
    trust_remote_code=True             # Required for custom dataset loading
)

print(dataset)
```

#### 🔹 Option 2: Download via [**Kaggle**](https://www.kaggle.com/datasets/dsliu0011/causal3d-image-dataset) + Croissant
```
import mlcroissant as mlc
import pandas as pd

# Load the dataset metadata from Kaggle
croissant_dataset = mlc.Dataset(
    "https://www.kaggle.com/datasets/dsliu0011/causal3d-image-dataset/croissant/download"
)

# List available record sets
record_sets = croissant_dataset.metadata.record_sets
print(record_sets)

# Load records from the first record set
df = pd.DataFrame(croissant_dataset.records(record_set=record_sets[0].uuid))
print(df.head())
```

---

## 📌 Overview

While recent progress in AI and computer vision has been remarkable, there remains a major gap in evaluating causal reasoning over complex visual inputs. **Causal3D** bridges this gap by providing:

- **19 curated 3D-scene datasets** simulating diverse real-world causal phenomena.
- Paired **tabular causal graphs** and **image observations** across multiple views and backgrounds.
- Benchmarks for evaluating models in both **structured** (tabular) and **unstructured** (image) modalities.

---

## 🧩 Dataset Structure

Each sub-dataset (scene) contains:


- `images/`: Rendered images under different camera views and backgrounds.
- `tabular.csv`: Instance-level annotations including object attributes in causal graph.


## 🖼️ Visual Previews

Below are example images from different Causal3D scenes:

<table>
  <tr>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/parabola.png" width="250"/><br/>parabola
    </td>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/convex.png" width="250"/><br/>convex
    </td>
  </tr>
  <tr>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/magnetic.png" width="200"/><br/>magnetic
    </td>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/pendulum.png" width="200"/><br/>pendulum
    </td>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/reflection.png" width="200"/><br/>reflection
    </td>
  </tr>
  <tr>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/seesaw.png" width="200"/><br/>seesaw
    </td>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/spring.png" width="200"/><br/>spring
    </td>
    <td align="center">
      <img src="https://huggingface.co/datasets/LLDDSS/Causal3D/resolve/main/preview/water_flow.png" width="200"/><br/>water_flow
    </td>
  </tr>
</table>

<!-- - `causal_graph.json`: Ground-truth causal structure (as adjacency matrix or graph).
- `view_info.json`: Camera/viewpoint metadata.
- `split.json`: Recommended train/val/test splits for benchmarking. -->

---

## 🎯 Evaluation Tasks

Causal3D supports a range of causal reasoning tasks, including:

- **Causal discovery** from image sequences or tables
- **Intervention prediction** under modified object states or backgrounds
- **Counterfactual reasoning** across views
- **VLM-based causal inference** given multimodal prompts

---

## 📊 Benchmark Results

We evaluate a diverse set of methods:

- **Classical causal discovery**: PC, GES, NOTEARS
- **Causal representation learning**: CausalVAE, ICM-based encoders
- **Vision-Language and Large Language Models**: GPT-4V, Claude-3.5, Gemini-1.5

**Key Findings**:

- As causal structures grow more complex, **model performance drops significantly** without strong prior assumptions.
- A noticeable performance gap exists between models trained on structured data and those applied directly to visual inputs.

---



<!-- ## 🔍 Example Use Case

```python
from causal3d import load_scene_data

scene = "SpringPendulum"
data = load_scene_data(scene, split="train")
images = data["images"]
metadata = data["table"]
graph = data["causal_graph"] -->