TextGNN: Improving Text Encoder via Graph Neural Network in Sponsored Search

TL;DR

TextGNN enhances text encoding for sponsored search by integrating user behavior graph information into twin tower models, leading to improved accuracy and revenue metrics while maintaining low latency.

Contribution

The paper introduces TextGNN, a novel graph neural network extension for twin tower text encoders that incorporates user behavior data to improve intent understanding.

Findings

0.14% increase in ROC-AUC offline

2.03% revenue increase online

2.32% decrease in ad defect rate online

Abstract

Text encoders based on C-DSSM or transformers have demonstrated strong performance in many Natural Language Processing (NLP) tasks. Low latency variants of these models have also been developed in recent years in order to apply them in the field of sponsored search which has strict computational constraints. However these models are not the panacea to solve all the Natural Language Understanding (NLU) challenges as the pure semantic information in the data is not sufficient to fully identify the user intents. We propose the TextGNN model that naturally extends the strong twin tower structured encoders with the complementary graph information from user historical behaviors, which serves as a natural guide to help us better understand the intents and hence generate better language representations. The model inherits all the benefits of twin tower models such as C-DSSM and TwinBERT so that it can still be used in the low latency environment while achieving a significant performance gain than the strong encoder-only counterpart baseline models in both offline evaluations and online production system. In offline experiments, the model achieves a 0.14% overall increase in ROC-AUC with a 1% increased accuracy for long-tail low-frequency Ads, and in the online A/B testing, the model shows a 2.03% increase in Revenue Per Mille with a 2.32% decrease in Ad defect rate.