VL-CheckList: Evaluating Pre-trained Vision-Language Models with Objects, Attributes and Relations
This provides a more detailed evaluation method for researchers and practitioners in vision-language AI, though it is incremental as it adapts an existing NLP testing approach to a new domain.
The authors tackled the problem of evaluating vision-language pretraining (VLP) models by proposing VL-CheckList, a framework that assesses models based on objects, attributes, and relations, revealing fine-grained differences among seven popular models that were not visible from downstream task accuracy alone.
Vision-Language Pretraining (VLP) models have recently successfully facilitated many cross-modal downstream tasks. Most existing works evaluated their systems by comparing the fine-tuned downstream task performance. However, only average downstream task accuracy provides little information about the pros and cons of each VLP method, let alone provides insights on how the community can improve the systems in the future. Inspired by the CheckList for testing natural language processing, we exploit VL-CheckList, a novel framework to understand the capabilities of VLP models. The proposed method divides the image-texting ability of a VLP model into three categories: objects, attributes, and relations, and uses a novel taxonomy to further break down these three aspects. We conduct comprehensive studies to analyze seven recently popular VLP models via the proposed framework. Results confirm the effectiveness of the proposed method by revealing fine-grained differences among the compared models that were not visible from downstream task-only evaluation. Further results show promising research direction in building better VLP models. Our data and code are available at: https://github.com/om-ai-lab/VL-CheckList.