Christoph Trattner gave a presentation on applying cognitive models to recommender systems. He discussed how ACT-R, an influential cognitive architecture, has been used to model temporal tagging behavior with power functions performing better than exponential functions. Experiments applying this approach achieved state-of-the-art results in predicting tag reuse and recommending tags. A two-step collaborative filtering approach that integrated tags and time (CIRTT) also outperformed baselines in recommending items. The presentation concluded with details on an open-source code framework for this research.