InterVLS: Interactive Model Understanding and Improvement with Vision-Language Surrogates
This addresses the need for more interpretable and accessible model understanding and improvement tools in critical applications, though it appears incremental as it builds on existing visual concept-based methods.
The paper tackles the problem of improving deep learning model interpretability and post-understanding model improvement by introducing InterVLS, a system that discovers text-aligned concepts and uses model-agnostic linear surrogates to measure influence, with a user study showing it effectively helps users identify influential concepts and adjust them to improve models.
Deep learning models are widely used in critical applications, highlighting the need for pre-deployment model understanding and improvement. Visual concept-based methods, while increasingly used for this purpose, face challenges: (1) most concepts lack interpretability, (2) existing methods require model knowledge, often unavailable at run time. Additionally, (3) there lacks a no-code method for post-understanding model improvement. Addressing these, we present InterVLS. The system facilitates model understanding by discovering text-aligned concepts, measuring their influence with model-agnostic linear surrogates. Employing visual analytics, InterVLS offers concept-based explanations and performance insights. It enables users to adjust concept influences to update a model, facilitating no-code model improvement. We evaluate InterVLS in a user study, illustrating its functionality with two scenarios. Results indicates that InterVLS is effective to help users identify influential concepts to a model, gain insights and adjust concept influence to improve the model. We conclude with a discussion based on our study results.