SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation

SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation

Scientific image generation faces challenges in semantic alignment and logical reasoning, prompting the creation of SciIR-82k dataset and SciIR-Bench evaluation framework to improv…

Hugging Face · Daily Papers ·Zhiyuan Ma, Zhengfeng Shi · ·▲ 3 upvotes

Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.

Autores: Zhiyuan Ma, Zhengfeng Shi, Yuning An, Peize Li, Jiabao Wei, Ruijie Li

  • 3 upvotes da comunidade
  • Temas: Text-to-Image, scientific image generation, Peirce's Semiotic Triad, Entity Structure, Scientific Process, Scientific Law

Resumo

Resumo original (em inglês), extraído do paper:

Scientific image generation faces challenges in semantic alignment and logical reasoning, prompting the creation of SciIR-82k dataset and SciIR-Bench evaluation framework to improve scientific reasoning capabilities in text-to-image models.

Onde ler

compartilhar: