Their 2015 paper on how they were doing recommenders also mentions that they used bits and pieces of the prize work: https://gwern.net/doc/reinforcement-learnin

Their 2015 paper on how they were doing recommenders also mentions that they used bits and pieces of the prize work: https://gwern.net/doc/reinforcement-learning/exploration/2015-gomezuribe.pdf https://gwern.net/doc/reinforcement-learning/exploration/201...