Retrieve, Caption, Generate: Visual Grounding for Enhancing Commonsense in Text Generation Models
This work addresses commonsense reasoning in text generation for AI applications, but it is incremental as it builds on existing Transformer models and tasks.
The authors tackled the problem of enhancing commonsense in text generation models by using multimodal information from images, resulting in noticeable improvements in performance, addressing issues like poor commonsense, fluency, and specificity in baseline generations.
We investigate the use of multimodal information contained in images as an effective method for enhancing the commonsense of Transformer models for text generation. We perform experiments using BART and T5 on concept-to-text generation, specifically the task of generative commonsense reasoning, or CommonGen. We call our approach VisCTG: Visually Grounded Concept-to-Text Generation. VisCTG involves captioning images representing appropriate everyday scenarios, and using these captions to enrich and steer the generation process. Comprehensive evaluation and analysis demonstrate that VisCTG noticeably improves model performance while successfully addressing several issues of the baseline generations, including poor commonsense, fluency, and specificity.