RaMen: Multi-Strategy Multi-Modal Learning for Bundle Construction

2025-07-18

AI Summary

Existing bundle recommendation methods primarily rely on bipartite graphs or semantically enhanced item representations, failing to adequately model the intricate explicit combinatorial logic and implicit user intents inherent in real-world bundles. To address this, we propose a multi-strategy fusion framework for bundle representation learning. First, we design an *explicit strategy-aware module* that captures structured item-level combinatorial relationships via hyperedge dependency modeling and multimodal attention. Second, we introduce an *implicit strategy-aware module* that uncovers latent collaborative intentions through hypergraph message passing and strategy-aligned discriminative learning, further unified via cross-strategy knowledge transfer. Extensive experiments on multiple real-world datasets demonstrate that our method significantly outperforms state-of-the-art baselines, validating its effectiveness and generalizability in complex bundle construction tasks.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Search and Optimization: Combinatorial Optimization
• Machine Learning: Multimodal Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Existing studies on bundle construction have relied merely on user feedback via bipartite graphs or enhanced item representations using semantic information. These approaches fail to capture elaborate relations hidden in real-world bundle structures, resulting in suboptimal bundle representations. To overcome this limitation, we propose RaMen, a novel method that provides a holistic multi-strategy approach for bundle construction. RaMen utilizes both intrinsic (characteristics) and extrinsic (collaborative signals) information to model bundle structures through Explicit Strategy-aware Learning (ESL) and Implicit Strategy-aware Learning (ISL). ESL employs task-specific attention mechanisms to encode multi-modal data and direct collaborative relations between items, thereby explicitly capturing essential bundle features. Moreover, ISL computes hyperedge dependencies and hypergraph message passing to uncover shared latent intents among groups of items. Integrating diverse strategies enables RaMen to learn more comprehensive and robust bundle representations. Meanwhile, Multi-strategy Alignment & Discrimination module is employed to facilitate knowledge transfer between learning strategies and ensure discrimination between items/bundles. Extensive experiments demonstrate the effectiveness of RaMen over state-of-the-art models on various domains, justifying valuable insights into complex item set problems.

Problem

• Capturing hidden relations in bundle structures
• Integrating intrinsic and extrinsic bundle information
• Learning comprehensive bundle representations multi-modally

Innovation

• Explicit Strategy-aware Learning with attention mechanisms
• Implicit Strategy-aware Learning via hypergraph processing
• Multi-strategy Alignment & Discrimination for knowledge transfer