TCNN: Triple Convolutional Neural Network Models for Retrieval-based Question Answering System in E-commerce

TL;DR

The paper tackles retrieval-based QA in e-commerce by unifying semantic matching of a user query with both the knowledge entry title and its answer. It introduces Triple Convolutional Neural Networks (TCNN) and two attention-enhanced variants (ATCNN-1, ATCNN-2) that process Q, T, and A with weight-sharing CNNs and attention signals to improve relatedness prediction. Experimental results show TCNN-based models outperform strong IR baselines, with ATCNN-2 achieving the best performance and maintaining low online latency (~1 ms), while ATCNN-1 underperforms due to mismatched relations between Q-T and Q-A. The work demonstrates the value of incorporating knowledge content beyond titles in retrieval-based QA and highlights practical viability for e-commerce systems.

Abstract

Automatic question-answering (QA) systems have boomed during last few years, and commonly used techniques can be roughly categorized into Information Retrieval (IR)-based and generation-based. A key solution to the IR based models is to retrieve the most similar knowledge entries of a given query from a QA knowledge base, and then rerank those knowledge entries with semantic matching models. In this paper, we aim to improve an IR based e-commerce QA system-AliMe with proposed text matching models, including a basic Triple Convolutional Neural Network (TCNN) model and two Attention-based TCNN (ATCNN) models. Experimental results show their effect.

Figures (3)

• Figure 1: Architecture of TCNN.
• Figure 2: One block in ATCNN-1.
• Figure 3: One block in ATCNN-2.