From Unstructured to Structured: LLM-Guided Attribute Graphs for Entity Search and Ranking

TL;DR

This paper introduces a two-stage method using LLMs to convert unstructured product data into structured attribute graphs, enhancing entity search and ranking accuracy in e-commerce without training data.

Contribution

It presents a novel approach combining LLM-driven attribute extraction and graph-based ranking, reducing token usage and improving zero-shot ranking performance.

Findings

Reduces per-product token usage by 57%.

Achieves over 5% improvement in average precision.

Outperforms multiple baselines in zero-shot scenarios.

Abstract

Entity search, i.e., finding the most similar entities to a query entity, faces unique challenges in e-commerce, where product similarity varies across categories and contexts. Traditional embedding-based approaches often struggle to capture nuanced context-specific attribute relevance. In this paper, we present a two-stage approach combining Large Language Model (LLM)-driven attribute graph construction with graph-aware LLM ranking. In the offline stage, we extract structured product attributes from unstructured text, and construct a reusable attribute graph with category-aware schemas. In the online stage, we rank retrieved candidates by reasoning over this structured representation rather than raw text, reducing per-product token usage by 57% while improving ranking precision. Experiments show that our approach outperforms multiple baselines under zero-shot scenarios, achieving a over 5% improvement in average precision without requiring training data, generalizes robustly across diverse product categories, and shows immense potential for real-world deployment.