Towards Multi-Behavior Multi-Task Recommendation via Behavior-informed Graph Embedding Learning

TL;DR

This work tackles multi-behavior multi-task recommendation (MMR) by introducing BiGEL, a behavior-informed graph embedding framework. BiGEL combines cascading GCN learning with cascading gated feedback, global context enhancement, and contrastive preference alignment to optimize both target and auxiliary behaviors and mitigate preference bias. The model jointly optimizes multiple behavior-specific BPR objectives while employing a contrastive loss to align target and global embeddings, and it demonstrates superior performance on two real-world datasets against strong baselines. The results highlight the value of integrating behavior-aware cascades with global context and contrastive alignment to improve personalized recommendations across behavior types, with practical implications for more balanced and robust multi-task recommender systems.

Abstract

Multi-behavior recommendation (MBR) aims to improve the performance w.r.t. the target behavior (i.e., purchase) by leveraging auxiliary behaviors (e.g., click, favourite). However, in real-world scenarios, a recommendation method often needs to process different types of behaviors and generate personalized lists for each task (i.e., each behavior type). Such a new recommendation problem is referred to as multi-behavior multi-task recommendation (MMR). So far, the most powerful MBR methods usually model multi-behavior interactions using a cascading graph paradigm. Although significant progress has been made in optimizing the performance of the target behavior, it often neglects the performance of auxiliary behaviors. To compensate for the deficiencies of the cascading paradigm, we propose a novel solution for MMR, i.e., behavior-informed graph embedding learning (BiGEL). Specifically, we first obtain a set of behavior-aware embeddings by using a cascading graph paradigm. Subsequently, we introduce three key modules to improve the performance of the model. The cascading gated feedback (CGF) module enables a feedback-driven optimization process by integrating feedback from the target behavior to refine the auxiliary behaviors preferences. The global context enhancement (GCE) module integrates the global context to maintain the user's overall preferences, preventing the loss of key preferences due to individual behavior graph modeling. Finally, the contrastive preference alignment (CPA) module addresses the potential changes in user preferences during the cascading process by aligning the preferences of the target behaviors with the global preferences through contrastive learning. Extensive experiments on two real-world datasets demonstrate the effectiveness of our BiGEL compared with ten very competitive methods.

Figures (7)

• Figure 1: Illustration of multi-behavior multi-task recommendation.
• Figure 2: Illustration of our BiGEL with three user behavior types as an example, i.e., click ($b_1$), favourite ($b_2$), and purchase ($b_3$). First, the cascading GCN learning (CGL) component captures the natural dependencies among different behaviors and models their sequential relationships. Then, the cascading gated feedback (CGF) module uses a gating mechanism to propagate the embedding of the target behavior ($b_3$) back to the preceding auxiliary behaviors ($b_1$, $b_2$), and the global context enhancement (GCE) module is designed to refine these behaviors. Next, the contrastive projection alignment (CPA) module aligns the target behavior embedding with the global embedding. Finally, we predict a user's preference scores w.r.t. each behavior type.
• Figure 3: Recommendation performance on JD (top) and UB (bottom) with different cutoff values K.
• Figure 4: Variation of the recommendation performance when removing GCE, CPA, CGF or their combinations.
• Figure 5: Impact of different MTL modules on our BiGEL.