Add task category and improve dataset card

README.md

@@ -1,4 +1,10 @@
 ---
+language:
+- en
+license: mit
+task_categories:
+- image-text-to-text
+arxiv: 2505.20310
 dataset_info:
   features:
   - name: field
@@ -34,25 +40,44 @@ configs:
     path: data/agriculture-*
   - split: environment
     path: data/environment-*
-license: mit
-language:
-- en
 ---
 
 # Manalyzer: End-to-end Automated Meta-analysis with Multi-agent System
 
+[**Project Page**](https://black-yt.github.io/meta-analysis-page/) | [**Paper**](https://huggingface.co/papers/2505.20310) | [**GitHub**](https://github.com/black-yt/Manalyzer)
+
 ## Overview
 
-Meta-analysis is a systematic research methodology that synthesizes data from multiple existing studies to derive comprehensive conclusions. This approach not only mitigates limitations inherent in individual studies but also facilitates novel discoveries through integrated data analysis. Traditional meta-analysis involves a complex multi-stage pipeline including literature retrieval, paper screening, and data extraction, which demands substantial human effort and time. However, while LLM-based methods can accelerate certain stages, they still face significant challenges, such as hallucinations in paper screening and data extraction. In this paper, we propose a multi-agent system, Manalyzer, which achieves end-to-end automated meta-analysis through tool calls. The hybrid review, hierarchical extraction, self-proving, and feedback checking strategies implemented in Manalyzer significantly alleviate these two hallucinations. To comprehensively evaluate the performance of meta-analysis, we construct a new benchmark comprising 729 papers across 3 domains, encompassing text, image, and table modalities, with over 10,000 data points. Extensive experiments demonstrate that Manalyzer achieves significant performance improvements over the LLM baseline in multi meta-analysis tasks.
+Meta-analysis is a systematic research methodology that synthesizes data from multiple existing studies to derive comprehensive conclusions. Traditional meta-analysis involves a complex multi-stage pipeline including literature retrieval, paper screening, and data extraction, which demands substantial human effort and time.
+
+**Manalyzer** is a multi-agent system that achieves end-to-end automated meta-analysis through tool calls. This repository contains the benchmark constructed to evaluate meta-analysis performance, comprising 729 papers across 3 domains (Atmosphere, Agriculture, and Environment), encompassing text, image, and table modalities, with over 10,000 data points.
+
+## Dataset Structure
 
-## Paper
+The benchmark consists of 729 papers across 3 scientific domains:
+- **Atmosphere**: 1,196 examples
+- **Agriculture**: 4,336 examples
+- **Environment**: 1,125 examples
 
-https://arxiv.org/pdf/2505.20310
+### Data Fields
 
-## Project Page
+Each example in the dataset contains:
+- `field`: The scientific domain (Atmosphere, Agriculture, or Environment).
+- `paper_idx`: Unique index of the source paper.
+- `doi`: Digital Object Identifier of the source paper.
+- `type`: Category of the data point.
+- `table_or_image`: Visual modality (extracted image of a table or figure).
+- `text_or_caption`: Associated text or caption providing context for the visual content.
 
-https://black-yt.github.io/meta-analysis-page/
+## Citation
 
-## GitHub Code
+If you find this dataset or the Manalyzer system useful in your research, please cite:
 
-https://github.com/black-yt/Manalyzer
+```bibtex
+@article{xu2025manalyzer,
+  title={Manalyzer: End-to-end Automated Meta-analysis with Multi-agent System},
+  author={Xu, Wanghan and Zhang, Wenlong and Ling, Fenghua and Fei, Ben and Hu, Yusong and Ren, Fangxuan and Lin, Jintai and Ouyang, Wanli and Bai, Lei},
+  journal={arXiv preprint arXiv:2505.20310},
+  year={2025}
+}
+```