The Deconfounded Recommender: A Causal Inference Approach to Recommendation

TL;DR

This paper introduces a causal inference framework for recommendation systems, addressing unobserved confounders to improve the accuracy and stability of recommendations by modeling user-movie interactions as causal treatments and outcomes.

Contribution

It develops the deconfounded recommender, a novel two-stage probabilistic model that accounts for unobserved confounders, enhancing causal prediction in recommendation systems.

Findings

Improves recommendation accuracy by removing confounding bias.

Achieves stable performance under intervention scenarios.

Effectively models causal effects in user-movie interactions.

Abstract

The goal of recommendation is to show users items that they will like. Though usually framed as a prediction, the spirit of recommendation is to answer an interventional question---for each user and movie, what would the rating be if we "forced" the user to watch the movie? To this end, we develop a causal approach to recommendation, one where watching a movie is a "treatment" and a user's rating is an "outcome." The problem is there may be unobserved confounders, variables that affect both which movies the users watch and how they rate them; unobserved confounders impede causal predictions with observational data. To solve this problem, we develop the deconfounded recommender, a way to use classical recommendation models for causal recommendation. Following Wang & Blei [23], the deconfounded recommender involves two probabilistic models. The first models which movies the users watch; it provides a substitute for the unobserved confounders. The second one models how each user rates each movie; it employs the substitute to help account for confounders. This two-stage approach removes bias due to confounding. It improves recommendation and enjoys stable performance against interventions on test sets.