EAVE: Efficient Product Attribute Value Extraction via Lightweight Sparse-layer Interaction

TL;DR

EAVE introduces a lightweight, efficient method for product attribute value extraction that balances performance and speed, especially suited for scenarios with many attributes and long contexts.

Contribution

The paper presents a novel sparse-layer interaction mechanism that enables efficient attribute value extraction by reusing heavy encodings and lightweight joint encoding.

Findings

Achieves significant efficiency improvements over existing methods.

Maintains comparable extraction accuracy with marginal performance loss.

Effective on benchmarks with long contexts and many attributes.

Abstract

Product attribute value extraction involves identifying the specific values associated with various attributes from a product profile. While existing methods often prioritize the development of effective models to improve extraction performance, there has been limited emphasis on extraction efficiency. However, in real-world scenarios, products are typically associated with multiple attributes, necessitating multiple extractions to obtain all corresponding values. In this work, we propose an Efficient product Attribute Value Extraction (EAVE) approach via lightweight sparse-layer interaction. Specifically, we employ a heavy encoder to separately encode the product context and attribute. The resulting non-interacting heavy representations of the context can be cached and reused for all attributes. Additionally, we introduce a light encoder to jointly encode the context and the attribute, facilitating lightweight interactions between them. To enrich the interaction within the lightweight encoder, we design a sparse-layer interaction module to fuse the non-interacting heavy representation into the lightweight encoder. Comprehensive evaluation on two benchmarks demonstrate that our method achieves significant efficiency gains with neutral or marginal loss in performance when the context is long and number of attributes is large. Our code is available \href{https://anonymous.4open.science/r/EAVE-EA18}{here}.