Image-guided topic modeling for interpretable privacy classification
This work addresses the need for interpretable privacy classification in images, which is crucial for users and developers in privacy-sensitive applications, though it is incremental as it builds on existing multimodal and interpretable methods.
The paper tackled the problem of predicting and explaining private information in images by proposing a method that uses natural language content descriptors with privacy scores, resulting in a classifier that outperforms the reference interpretable method by 5 percentage points in accuracy and matches non-interpretable state-of-the-art performance.
Predicting and explaining the private information contained in an image in human-understandable terms is a complex and contextual task. This task is challenging even for large language models. To facilitate the understanding of privacy decisions, we propose to predict image privacy based on a set of natural language content descriptors. These content descriptors are associated with privacy scores that reflect how people perceive image content. We generate descriptors with our novel Image-guided Topic Modeling (ITM) approach. ITM leverages, via multimodality alignment, both vision information and image textual descriptions from a vision language model. We use the ITM-generated descriptors to learn a privacy predictor, Priv$\times$ITM, whose decisions are interpretable by design. Our Priv$\times$ITM classifier outperforms the reference interpretable method by 5 percentage points in accuracy and performs comparably to the current non-interpretable state-of-the-art model.