Expressive recommender systems through normalized nonnegative models
This addresses the need for interpretable and efficient recommender systems in domains like movies or documents, though it appears incremental as it builds on existing probability theory models.
The paper tackles the problem of creating expressive recommender systems by introducing normalized nonnegative models (NNM), which achieve high predictive power, computational tractability, and expressive representations of users and items.
We introduce normalized nonnegative models (NNM) for explorative data analysis. NNMs are partial convexifications of models from probability theory. We demonstrate their value at the example of item recommendation. We show that NNM-based recommender systems satisfy three criteria that all recommender systems should ideally satisfy: high predictive power, computational tractability, and expressive representations of users and items. Expressive user and item representations are important in practice to succinctly summarize the pool of customers and the pool of items. In NNMs, user representations are expressive because each user's preference can be regarded as normalized mixture of preferences of stereotypical users. The interpretability of item and user representations allow us to arrange properties of items (e.g., genres of movies or topics of documents) or users (e.g., personality traits) hierarchically.