Search Solutions 2020 – Tutorial Report

On 24th November 2020, a day before the Search Solutions main event, a Search Solutions tutorial on Reinforcement Learning for Information Retrieval was given by Alexander Kuhnle, Miguel Aroca-Ouellette, John Reid and Dell Zhang from Blue Prism AI Labs. Reinforcement Learning (RL), as a third machine learning paradigm besides supervised and unsupervised learning, is gaining more and more attraction in the Information Retrieval (IR) community, in research as well as industry. After an introduction to RL and necessary basics, some more specific RL topics like Q-Learning with Deep Learning, policy gradient methods, actor-critics, etc were discussed. Finally, IR applications utilising RL were presented. The taught theory of this very well prepared and delivered tutorial was complemented by practical Python exercises with Google Colab. The tutorial attracted at peak times around 35 delegates. It was very well received and the feedback from the audiences at the closing session was positive. All material is available on the tutorial Website, which also contains additional resources.

Ingo Frommholz and Haiming Liu

About Ingo Frommholz
Ingo Frommholz

Ingo is Reader in Data Science at the University of Wolverhampton, UK. His research focuses on different aspects of information retrieval, in particular formal models for user-oriented search and digital libraries, with emphasis on polyrepresentation, annotation-based retrieval, interactive quantum-based IR and probabilistic and logic-based models. Ingo is actively sharing his passion for information retrieval and digital libraries on Twitter where he is known as @iFromm.

Leave a Reply

You must be logged in to post a comment.