AutoPKG: An Automated Framework for Dynamic E-commerce Product-Attribute Knowledge Graph Construction

2026-04-18

AI Summary

This work addresses the challenges of inconsistency, incompleteness, and high maintenance costs in e-commerce product attribute knowledge graphs by proposing an automated construction framework based on multi-agent large language models. The framework integrates multimodal product content to dynamically induce product types and attribute keys, collaboratively extracts attribute values from both text and images, and employs a central decision-making agent to ensure global graph consistency. It innovatively introduces a dynamic multi-agent mechanism supporting ontology evolution and a novel evaluation protocol tailored for dynamic product knowledge graphs. Evaluated on real-world Lazada data, the approach achieves a type-level WKE of 0.953 and an edge-level F1 score of 0.531; it further improves edge-level F1 by 0.152 and attribute extraction accuracy by 0.208 across three public benchmarks. Online A/B tests demonstrate GMV gains of 3.81%, 5.32%, and 7.89% in badge, search, and recommendation scenarios, respectively.

Technology Category

• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Machine Learning: Graph-based Machine Learning
• Multiagent Systems: Mechanism Design

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Product attribute extraction in e-commerce is bottlenecked by ontologies that are inconsistent, incomplete, and costly to maintain. We present AutoPKG, a multi-agent Large Language Model (LLM) framework that automatically constructs a Product-attribute Knowledge Graph (PKG) from multimodal product content. AutoPKG induces product types and type-specific attribute keys on demand, extracts attribute values from text and images, and consolidates updates through a centralized decision agent that maintains a globally consistent canonical graph. We also propose an evaluation protocol for dynamic PKGs that measures type and key validity, consolidation quality, and edge-level accuracy for value assertions after canonicalization. On a large real-world marketplace catalog dataset from Lazada (Alibaba), AutoPKG achieves up to 0.953 Weighted Knowledge Efficiency (WKE) for product types, 0.724 WKE for attribute keys, and 0.531 edge-level F1 for multimodal value extraction. Across three public benchmarks, our method improves edge-level exact-match F1 by 0.152 and yields a precision gain of 0.208 on the attribute extraction application. Online A/B tests show that AutoPKG-derived attributes increase Gross Merchandise Value (GMV) in Badge by 3.81 percent, in Search by 5.32 percent, and in Recommendation by 7.89 percent, supporting the practical value of AutoPKG in production.

Problem

• product attribute extraction
• knowledge graph construction
• ontology maintenance
• e-commerce
• multimodal content

Innovation

• multi-agent LLM
• dynamic knowledge graph
• multimodal attribute extraction
• canonicalization
• product-attribute ontology