Diversity Recommendation via Causal Deconfounding of Co-purchase Relations and Counterfactual Exposure

TL;DR

This paper introduces Cadence, a causal deconfounding framework for item-to-item recommendation that improves diversity without sacrificing accuracy by removing popularity bias and simulating high-exposure scenarios.

Contribution

The paper proposes a novel causal deconfounding approach using UACR and counterfactual exposure to enhance recommendation diversity while maintaining relevance.

Findings

Outperforms state-of-the-art models in diversity and accuracy

Effectively removes popularity bias from item relationships

Demonstrates transferability and efficiency on real datasets

Abstract

Beyond user-item modeling, item-to-item relationships are increasingly used to enhance recommendation. However, common methods largely rely on co-occurrence, making them prone to item popularity bias and user attributes, which degrades embedding quality and performance. Meanwhile, although diversity is acknowledged as a key aspect of recommendation quality, existing research offers limited attention to it, with a notable lack of causal perspectives and theoretical grounding. To address these challenges, we propose Cadence: Diversity Recommendation via Causal Deconfounding of Co-purchase Relations and Counterfactual Exposure - a plug-and-play framework built upon LightGCN as the backbone, primarily designed to enhance recommendation diversity while preserving accuracy. First, we compute the Unbiased Asymmetric Co-purchase Relationship (UACR) between items - excluding item popularity and user attributes - to construct a deconfounded directed item graph, with an aggregation mechanism to refine embeddings. Second, we leverage UACR to identify diverse categories of items that exhibit strong causal relevance to a user's interacted items but have not yet been engaged with. We then simulate their behavior under high-exposure scenarios, thereby significantly enhancing recommendation diversity while preserving relevance. Extensive experiments on real-world datasets demonstrate that our method consistently outperforms state-of-the-art diversity models in both diversity and accuracy, and further validates its effectiveness, transferability, and efficiency over baselines.