Model-agnostic post-hoc explainability for recommender systems
This work addresses the need for transparency in recommender systems for users and developers, but it is incremental as it adapts existing diagnostic techniques to a new domain.
The research tackled the problem of interpretability in recommender systems by developing a model-agnostic post-hoc explainability method based on deletion diagnostics, which quantifies the influence of users or items on recommendations and was demonstrated on Neural Collaborative Filtering and Singular Value Decomposition across MovieLens and Amazon Reviews datasets.
Recommender systems often benefit from complex feature embeddings and deep learning algorithms, which deliver sophisticated recommendations that enhance user experience, engagement, and revenue. However, these methods frequently reduce the interpretability and transparency of the system. In this research, we develop a systematic application, adaptation, and evaluation of deletion diagnostics in the recommender setting. The method compares the performance of a model to that of a similar model trained without a specific user or item, allowing us to quantify how that observation influences the recommender, either positively or negatively. To demonstrate its model-agnostic nature, the proposal is applied to both Neural Collaborative Filtering (NCF), a widely used deep learning-based recommender, and Singular Value Decomposition (SVD), a classical collaborative filtering technique. Experiments on the MovieLens and Amazon Reviews datasets provide insights into model behavior and highlight the generality of the approach across different recommendation paradigms.