FigCaps-HF: A Figure-to-Caption Generative Framework and Benchmark with Human Feedback
This addresses the issue of misaligned captions in scientific documents, which is incremental as it builds on existing captioning methods with a focus on human feedback.
The paper tackles the problem of generating high-quality captions for scientific figures by introducing FigCaps-HF, a framework that incorporates domain expert feedback to optimize for reader preferences, resulting in improvements such as a 35.7% gain in ROUGE when using BLIP as the base model.
Captions are crucial for understanding scientific visualizations and documents. Existing captioning methods for scientific figures rely on figure-caption pairs extracted from documents for training, many of which fall short with respect to metrics like helpfulness, explainability, and visual-descriptiveness [15] leading to generated captions being misaligned with reader preferences. To enable the generation of high-quality figure captions, we introduce FigCaps-HF a new framework for figure-caption generation that can incorporate domain expert feedback in generating captions optimized for reader preferences. Our framework comprises of 1) an automatic method for evaluating quality of figure-caption pairs, 2) a novel reinforcement learning with human feedback (RLHF) method to optimize a generative figure-to-caption model for reader preferences. We demonstrate the effectiveness of our simple learning framework by improving performance over standard fine-tuning across different types of models. In particular, when using BLIP as the base model, our RLHF framework achieves a mean gain of 35.7%, 16.9%, and 9% in ROUGE, BLEU, and Meteor, respectively. Finally, we release a large-scale benchmark dataset with human feedback on figure-caption pairs to enable further evaluation and development of RLHF techniques for this problem.