DHA: Product Title Generation with Discriminative Hierarchical Attention for E-commerce

Zhu，Wenya1; Zhang，Yinghua2; Zhang，Yu3; Zhou，Yuhang1; Feng，Yinfu1; Wu，Yuxiang4; Da，Qing1; Zeng，Anxiang5

10.1007/978-3-031-05981-0_22

2022

26th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD)

0302-9743

1611-3349

978-3-031-05980-3

LECT NOTES ARTIF INT

13282 LNAI

275-287

MAY 16-19, 2022

SW Jiaotong Univ,Chengdu,PEOPLES R CHINA

GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND

SPRINGER INTERNATIONAL PUBLISHING AG

Product titles play an important role in E-Commerce sites. However, manually crafting product titles needs tremendous time and human effort. It is expected that product titles can be automatically generated, but existing generation methods usually require densely labeled data that are unavailable in the real world. We formulate a novel product title generation task that generates the title from the image and auxiliary information (e.g., category) to address the gap. To generate titles that are consistent with search queries, we construct the first large-scale dataset (AEPro) and propose a Discriminative Hierarchical Attention (DHA) model. The DHA model first identifies the image regions related to the product of interest (POI) with a POI attention module. Then, based on the title context, the identified image regions are further revised by a generation attention module. Finally, the titles are generated by dynamically attending to these image regions. Experiments on the AEPro dataset demonstrate the effectiveness of the DHA model. Besides, online A/B testing results show that 61.8 % of the titles generated by the DHA model are accepted directly or with minor modifications. The exposure rate of the products with machine-generated titles is improved by 40 %.

Discriminative attention Hierarchical attention Natural language generation

其他

英语

EI ; CPCI-S

Computer Science

Computer Science, Artificial Intelligence ; Computer Science, Information Systems

WOS:000870728100022

20222112138297

Large dataset ; Natural language processing systems

Data Processing and Image Processing:723.2 ; Computer Applications:723.5

2-s2.0-85130239113

Scopus

1.Alibaba,Hangzhou,China
2.Hong Kong University of Science and Technology,Kowloon,Hong Kong
3.Southern University of Science and Technology,Shenzhen,China
4.University College London,London,United Kingdom
5.Nanyang Technological University,Singapore,Singapore

Zhu，Wenya,Zhang，Yinghua,Zhang，Yu,et al. DHA: Product Title Generation with Discriminative Hierarchical Attention for E-commerce[C]. GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND:SPRINGER INTERNATIONAL PUBLISHING AG,2022:275-287.