README.md

metadata

dataset_info:
  features:
    - name: conference
      dtype: string
    - name: year
      dtype: int32
    - name: paper_id
      dtype: int32
    - name: title
      dtype: string
    - name: abstract
      dtype: string
    - name: topics
      sequence: string
    - name: image_url
      dtype: string
  splits:
    - name: train
      num_bytes: 15394703
      num_examples: 10305
    - name: validation
      num_bytes: 4461536
      num_examples: 3000
    - name: test
      num_bytes: 4464840
      num_examples: 3000
  download_size: 12550503
  dataset_size: 24321079
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*

POSTERSUM Dataset

Dataset Summary

The POSTERSUM dataset is a multimodal benchmark designed for the summarization of scientific posters into research paper abstracts. The dataset consists of 16,305 research posters collected from major machine learning conferences, including ICLR, ICML, and NeurIPS, spanning the years 2022-2024. Each poster is provided in image format along with its corresponding abstract as a summary. This dataset is intended for research in multimodal understanding and summarization tasks, particularly in vision-language models (VLMs) and Multimodal Large Language Models (MLLMs).

Dataset Details

Data Fields

Each record in the dataset contains the following fields:

• conference (string): Name of the conference where the research poster was presented (e.g., ICLR, ICML, NeurIPS).
• year (int): The year of the conference.
• paper_id (int): Conference identifier for the research paper associated with the poster.
• title (string): The title of the research paper.
• abstract (string): The human-written abstract of the paper, serving as the ground-truth summary for the poster.
• topics (list of strings): Machine learning topics related to the research (e.g., Reinforcement Learning, Natural Language Processing, Graph Neural Networks).
• image_url (string): URL to the image file of the scientific poster.

Dataset Statistics

• Total number of poster-summary pairs: 16,305
• Total number of unique topics: 137
• Average summary length: 224 tokens
• Train/Validation/Test split: 10,305 / 3,000 / 3,000

Citation

@misc{saxena2025postersummultimodalbenchmarkscientific,
      title={PosterSum: A Multimodal Benchmark for Scientific Poster Summarization}, 
      author={Rohit Saxena and Pasquale Minervini and Frank Keller},
      year={2025},
      eprint={2502.17540},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2502.17540}, 
}