Understanding Latent Factors Using a GWAP
This work addresses the interpretability issue in recommender systems for users, but it is incremental as it builds on existing latent factor models with a new descriptive method.
The paper tackled the problem of making latent factor models in recommender systems more interpretable by developing a game to generate semantic descriptions for factors, and a user study confirmed that the collected descriptions accurately reflected real-world characteristics.
Recommender systems relying on latent factor models often appear as black boxes to their users. Semantic descriptions for the factors might help to mitigate this problem. Achieving this automatically is, however, a non-straightforward task due to the models' statistical nature. We present an output-agreement game that represents factors by means of sample items and motivates players to create such descriptions. A user study shows that the collected output actually reflects real-world characteristics of the factors.