Abstract:A well-developed recommendation system can not only leverage multi-typed interactions (such as page view, add-to-cart, and purchase) to better identify user preferences, but also demonstrate high performance, low complexity, and strong interpretability. However, many existing solutions for multi-behavior recommendation fall short of intuitive modeling of real-world scenarios, leading to overly complex models with massive parameters and cumbersome components. In particular, they share two critical limitations: (1) Some pioneering models are built upon the strict assumption of cascade effects across behaviors, which contradicts multifarious behavior paths in practical applications. (2) Existing approaches fail to explicitly capture the unique idiosyncrasies of users and even neglect the inherent nature of items involved in the multi-behavior interactions. To this end, we propose a novel Directed Acyclic Graph Convolutional Network (DA-GCN) for the multi-behavior recommendation task. Specifically, we pinpoint the partial order relations within the monotonic behavior chain and extend it to personalized directed acyclic behavior graphs to exploit behavior dependencies. Then, a GCN-based directed edge encoder is employed to distill rich collaborative signals embodied by each directed edge. In light of the information flows over the directed acyclic structure, we propose an attentive aggregation module to gather messages from all potential antecedent behaviors, representing distinct perspectives to understand the terminated behavior. Thus, we obtain comprehensive representations for the follow-up behavior through learnable distributions over its preceding behaviors, explicitly reflecting personalized interactive patterns of users and underlying properties of items simultaneously. Finally, we design a customized multi-task learning objective for flexible joint optimization. Extensive experiments on public benchmarking datasets fully demonstrate the superiority of DA-GCN with significant performance improvement and computational efficiency over a wide range of state-of-the-art methods. Our code is available at https://github.com/xizhu1022/DA-GCN.

Learning Personalized Itemset Mapping for Cross-Domain Recommendation

SGLCMR: Self-supervised Graph Learning of Generalized Representations for Cross-Market Recommendation.

User Distribution Mapping Modelling with Collaborative Filtering for Cross Domain Recommendation

Category-aware Self-Supervised Graph Neural Network for Session-Based Recommendation

Deep Graph Mutual Learning for Cross-domain Recommendation

Time Interval-Enhanced Graph Neural Network for Shared-Account Cross-Domain Sequential Recommendation

Intra and Inter Domain HyperGraph Convolutional Network for Cross-Domain Recommendation

Gromov-Wasserstein Guided Representation Learning for Cross-Domain Recommendation.

Cross-Domain Sequential Recommendation via Neural Process

DDTCDR: Deep Dual Transfer Cross Domain Recommendation

DDGHM: Dual Dynamic Graph with Hybrid Metric Training for Cross-Domain Sequential Recommendation

Multi-Behavior Recommendation with Personalized Directed Acyclic Behavior Graphs

Joint Internal Multi-Interest Exploration and External Domain Alignment for Cross Domain Sequential Recommendation

A Collaborative Transfer Learning Framework for Cross-domain Recommendation

Deep Transfer Collaborative Filtering with Geometric Structure Preservation for Cross-Domain Recommendation

Gmr-Rec: Graph Mutual Regularization Learning for Multi-Domain Recommendation

Knowledge-aware Neural Collective Matrix Factorization for Cross-domain Recommendation

DA-GCN: A Domain-aware Attentive Graph Convolution Network for Shared-account Cross-domain Sequential Recommendation

One for All, All for One: Learning and Transferring User Embeddings for Cross-Domain Recommendation

Transferring Causal Mechanism over Meta-representations for Target-unknown Cross-domain Recommendation

Extracting latently overlapping users by graph neural network for non-overlapping cross-domain recommendation