Dual-Enhancement Product Bundling: Bridging Interactive Graph and Large Language Model

TL;DR

This paper introduces a dual-enhancement approach combining interactive graph learning and large language models to improve product bundling, addressing cold-start issues and modeling limitations.

Contribution

It proposes a novel graph-to-text paradigm with a Dynamic Concept Binding Mechanism to enhance LLM understanding of product graphs for bundling.

Findings

Achieved 6.3%-26.5% improvements over baselines on three benchmarks.

Effectively models combinatorial constraints via graph-to-text translation.

Bridges the gap between graph-based and language-based product understanding.

Abstract

Product bundling boosts e-commerce revenue by recommending complementary item combinations. However, existing methods face two critical challenges: (1) collaborative filtering approaches struggle with cold-start items owing to dependency on historical interactions, and (2) LLMs lack inherent capability to model interactive graph directly. To bridge this gap, we propose a dual-enhancement method that integrates interactive graph learning and LLM-based semantic understanding for product bundling. Our method introduces a graph-to-text paradigm, which leverages a Dynamic Concept Binding Mechanism (DCBM) to translate graph structures into natural language prompts. The DCBM plays a critical role in aligning domain-specific entities with LLM tokenization, enabling effective comprehension of combinatorial constraints. Experiments on three benchmarks (POG, POG_dense, Steam) demonstrate 6.3%-26.5% improvements over state-of-the-art baselines.