README.md

---
license: apache-2.0
paperswithcode_id: marvel
pretty_name: MARVEL (Multidimensional Abstraction and Reasoning through Visual Evaluation and Learning)
task_categories:
  - visual-question-answering
  - question-answering
  - multiple-choice
  - image-classification
task_ids:
  - multiple-choice-qa
  - closed-domain-qa
  - open-domain-qa
  - visual-question-answering
tags:
  - multi-modal-qa
  - geometry-qa
  - abstract-reasoning
  - geometry-reasoning
  - visual-puzzle
  - non-verbal-reasoning
  - abstract-shapes
language:
- en
size_categories:
- n<1K
configs:
  - config_name: default
    data_files: marvel.parquet
dataset_info:
  - config_name: default
    features:
      - name: id
        dtype: int64
      - name: pattern
        dtype: string
      - name: task_configuration
        dtype: string
      - name: avr_question
        dtype: string
      - name: explanation
        dtype: string
      - name: answer
        dtype: int64
      - name: f_perception_question
        dtype: string
      - name: f_perception_answer
        dtype: string
      - name: f_perception_distractor
        dtype: string
      - name: c_perception_question_tuple
        sequence: string
      - name: c_perception_answer_tuple
        sequence: string
      - name: file
        dtype: string
      - name: image
        dtype: image
---
## Dataset Details

### Dataset Description

MARVEL is a new comprehensive benchmark dataset that evaluates multi-modal large language models' abstract reasoning abilities in six patterns across five different task configurations, revealing significant performance gaps between humans and SoTA MLLMs.
![image](./marvel_illustration.jpeg)

### Dataset Sources [optional]

- **Repository:** https://github.com/1171-jpg/MARVEL_AVR
- **Paper [optional]:** https://arxiv.org/abs/2404.13591
- **Demo [optional]:** https://marvel770.github.io/

## Uses

Evaluations for multi-modal large language models' abstract reasoning abilities.

## Dataset Structure

The directory **images** keeps all images, and the file **marvel_labels.jsonl** provides annotations and explanations for all questions.

### Fields

- **id** is of ID of the question
- **pattern** is the high-level pattern category of the question
- **task_configuration** is the task configuration of the question
- **avr_question** is the text of the AVR question
- **answer** is the answer to the AVR question
- **explanation** is the textual reasoning process to answer the question
- **f_perception_question** is the fine-grained perception question
- **f_perception_answer** is the answer to the fine-grained perception question
- **f_perception_distractor** is the distractor of the fine-grained perception question
- **c_perception_question_tuple** is a list of coarse-grained perception questions
- **c_perception_answer_tuple** is a list of answers to the coarse-grained perception questions
- **file** is the path to the image of the question

## Citation [optional]

**BibTeX:**
```
@article{jiang2024marvel,
  title={MARVEL: Multidimensional Abstraction and Reasoning through Visual Evaluation and Learning},
  author={Jiang, Yifan and Zhang, Jiarui and Sun, Kexuan and Sourati, Zhivar and Ahrabian, Kian and Ma, Kaixin and Ilievski, Filip and Pujara, Jay},
  journal={arXiv preprint arXiv:2404.13591},
  year={2024}
}
```