Related papers: Towards Better Query Classification with Multi-Expert Knowledge Condensation in JD Ads Search

Towards Better Query Classification with Multi-Expert Knowledge Condensation in JD Ads Search

URL: http://arxiv.org/abs/2308.01098v3
Date: Sun, 19 Nov 2023 14:42:39 GMT
Title: Towards Better Query Classification with Multi-Expert Knowledge Condensation in JD Ads Search
Authors: Kun-Peng Ning, Ming Pang, Zheng Fang, Xue Jiang, Xi-Wei Zhao, Chang-Ping Peng, Zhan-Gang Lin, Jing-He Hu, Jing-Ping Shao
Abstract summary: shallow model FastText is widely used for efficient online inference. BERT is an effective solution, but it will cause a higher online inference latency and more expensive computing costs. We propose knowledge condensation to boost the classification performance of the online FastText model under strict low latency constraints.
Score: 12.701416688678622
License: http://creativecommons.org/licenses/by/4.0/
Abstract: Search query classification, as an effective way to understand user intents, is of great importance in real-world online ads systems. To ensure a lower latency, a shallow model (e.g. FastText) is widely used for efficient online inference. However, the representation ability of the FastText model is insufficient, resulting in poor classification performance, especially on some low-frequency queries and tailed categories. Using a deeper and more complex model (e.g. BERT) is an effective solution, but it will cause a higher online inference latency and more expensive computing costs. Thus, how to juggle both inference efficiency and classification performance is obviously of great practical importance. To overcome this challenge, in this paper, we propose knowledge condensation (KC), a simple yet effective knowledge distillation framework to boost the classification performance of the online FastText model under strict low latency constraints. Specifically, we propose to train an offline BERT model to retrieve more potentially relevant data. Benefiting from its powerful semantic representation, more relevant labels not exposed in the historical data will be added into the training set for better FastText model training. Moreover, a novel distribution-diverse multi-expert learning strategy is proposed to further improve the mining ability of relevant data. By training multiple BERT models from different data distributions, it can respectively perform better at high, middle, and low-frequency search queries. The model ensemble from multi-distribution makes its retrieval ability more powerful. We have deployed two versions of this framework in JD search, and both offline experiments and online A/B testing from multiple datasets have validated the effectiveness of the proposed approach.