Reinforcement Online Learning to Rank with Unbiased Reward Shaping

Abstract

Online learning to rank (OLTR) aims to learn a ranker directly from implicit feedback derived from users' interactions, such as clicks. Clicks however are a biased signal: specifically, top-ranked documents are likely to attract more clicks than documents down the ranking (position bias). In this paper, we propose a novel learning algorithm for OLTR that uses reinforcement learning to optimize rankers: Reinforcement Online Learning to Rank (ROLTR). In ROLTR, the gradients of the ranker are estimated based on the rewards assigned to clicked and unclicked documents. In order to de-bias the users' position bias contained in the reward signals, we introduce unbiased reward shaping functions that exploit inverse propensity scoring for clicked and unclicked documents. The fact that our method can also model unclicked documents provides a further advantage in that less users interactions are required to effectively train a ranker, thus providing gains in efficiency. Empirical evaluation on standard OLTR datasets shows that ROLTR achieves state-ofthe-art performance, and provides significantly better user experience than other OLTR approaches. To facilitate the reproducibility of our experiments, we make all experiment code available at https://github.com/ielab/OLTR.

Implicit Feedback for Dense Passage Retrieval: A Counterfactual Approach

“…This is because it is not straightforward to ascertain why a user has not clicked on a passage: is it because the passage is not relevant or because the user did not examine the passage? Nevertheless, we realize that there have been recent attempts to de-bias unclick signals that have used IPS-like algorithms [43,49]: these can potentially be integrated into our CoRocchio algorithm to further leverage the user's negative implicit feedback for dense retrievers. We leave this to future exploration.…”

“…To fully control users' biases and noise so that algorithms can be tested under different, controllable conditions, it is common for research in online LTR and counterfactual LTR to simulate users' clicks by relying on the relevance labels recorded in offline datasets [16,19,37,39,49,50]. Following this practice, we use a click model to simulate click behaviour and create a synthetic click log.…”