ConceptX: A Framework for Latent Concept Analysis
This addresses the problem of explainability in AI for users needing transparency, though it appears incremental as it builds on existing methods for concept analysis.
The paper tackles the opacity of deep neural networks by introducing ConceptX, a human-in-the-loop framework for interpreting and annotating latent concepts in pre-trained language models, enabling the discovery of concepts and generation of explanations through a graphical interface.
The opacity of deep neural networks remains a challenge in deploying solutions where explanation is as important as precision. We present ConceptX, a human-in-the-loop framework for interpreting and annotating latent representational space in pre-trained Language Models (pLMs). We use an unsupervised method to discover concepts learned in these models and enable a graphical interface for humans to generate explanations for the concepts. To facilitate the process, we provide auto-annotations of the concepts (based on traditional linguistic ontologies). Such annotations enable development of a linguistic resource that directly represents latent concepts learned within deep NLP models. These include not just traditional linguistic concepts, but also task-specific or sensitive concepts (words grouped based on gender or religious connotation) that helps the annotators to mark bias in the model. The framework consists of two parts (i) concept discovery and (ii) annotation platform.