Towards explainable evaluation of language models on the semantic similarity of visual concepts
This work addresses explainability and robustness issues in evaluation strategies for language models, focusing on visual vocabularies, but it is incremental as it builds on existing methods without introducing a new paradigm.
The authors tackled the problem of evaluating language models on semantic similarity for visual concepts by proposing explainable metrics that reveal the limitations of existing approaches and exposing vulnerabilities through adversarial interventions.
Recent breakthroughs in NLP research, such as the advent of Transformer models have indisputably contributed to major advancements in several tasks. However, few works research robustness and explainability issues of their evaluation strategies. In this work, we examine the behavior of high-performing pre-trained language models, focusing on the task of semantic similarity for visual vocabularies. First, we address the need for explainable evaluation metrics, necessary for understanding the conceptual quality of retrieved instances. Our proposed metrics provide valuable insights in local and global level, showcasing the inabilities of widely used approaches. Secondly, adversarial interventions on salient query semantics expose vulnerabilities of opaque metrics and highlight patterns in learned linguistic representations.