Safe Deployment for Counterfactual Learning to Rank by Shashank Gupta - ShareChat ML Seminar
The ShareChat Machine Learning Seminar invites world-leading researchers to present recent advances in their work in all areas related to machine learning and the problems we solve at ShareChat.
Abstract:
Counterfactual learning to rank (CLTR) relies on exposure-based inverse propensity scoring (IPS), an LTR-specific adaptation of IPS to correct position bias. While IPS can provide unbiased and consistent estimates, it often suffers from high variance. Especially when little click data is available, this variance can cause CLTR to learn sub-optimal ranking behaviour. Consequently, existing CLTR methods bring significant risks with them, as naively deploying their models can result in very negative user experiences. We introduce a novel risk-aware CLTR method with theoretical guarantees for safe deployment. We apply a novel exposure-based concept of risk regularization to IPS estimation for LTR. Our risk regularization penalizes the mismatch between the ranking behaviour of a learned model and a given safe model. Thereby, it ensures that learned ranking models stay close to a trusted model when there is high uncertainty in IPS estimation, which greatly reduces the risks during deployment. For the CLTR field, our novel exposure-based risk minimization method enables practitioners to adopt CLTR methods in a safer manner that mitigates many of the risks attached to previous methods.
Short Bio:
Shashank Gupta is a third-year PhD student at the University of Amsterdam (UvA) working under the supervision of Prof. Dr. Maarten de Rijke and Dr. Harrie Oosterhuis. His main research interests are counterfactual evaluation and learning from user interactions. Before joining UvA, he worked as a data scientist with the search team at Flipkart, an Indian e-commerce company. At Flipkart, he worked on various applied IR problems, such as personalized search, spell correction, and session length prediction. He completed his MS at IIIT-Hyderabad, India. For more details about his work, you can visit: http://shashank-gupta.com
#ShareChatMLSeminar #LifeAtShareChat