Back to Newsroom
Back to Newsroom

DeepBlue Technology comes out top at SIGIR, improving fundamental research results for e-commerce search.

Monday, 29 July 2019 02:20 AM

PARIS, FRANCE / ACCESSWIRE / July 29, 2019 / July 24, 2019: DeepBlue Technology emerged Champion in the High Precision Recall Task Challenge Competition at SIGIR 2019, organized by eBay. SIGIR, short for Special Interest Group on Information Retrieval, is the premier international forum for the presentation of new research results and for the demonstration of new systems and techniques in information retrieval. The annual international SIGIR conference, which began in 1978, is considered the most important in the field of information retrieval. The 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval took place on July 21-25, 2019 in Paris. The conference is backed up by the French Association for Information Retrieval and Applications (ARIA) which organizes the yearly IR French CORIA conference.

DeepBlue Technology, Monday, July 29, 2019, Press release picture

According to the website, this competition ‘targets a common problem in e-commerce search: identifying the items to show when sorted by non-relevance sorts.’ This is because users of

e-commerce search applications are often sorted by dimensions other than relevance. Though seemingly simple, achieving high relevancy rates while sorting non-relevance e-commerce searches is very tough because the search engines needs to perform higher order searches while taking typical search challenges into account like ambiguity and multiple query intents.

DeepBlue achieved pole position in a team of 20 competing teams, over 3 phases. DeepBlue’s team leader, Prof Wang Xinlei, said, ‘In the world of integrated technologies, commerce is going to be a natural extension of our capability to search on demand. We are interested to understand how users interact with platforms and how commerce and searches can be organically done to bring a better experience to users.’

In this competition, DeepBlue Technology used a methodology called Enhanced LSTM for Natural Language Inference, or ESIM for short. Using a gated network on top of ESIM, it allowed for ten times higher relevance between items presented in searches, even though the search might not be sorted according to relevance.

Such research efforts are integral to DeepBlue’s mission of serving humanity through AI, as it will be translated into features and commercialized in the company’s products, such as the quiXmart line of products and software.

For media contact, please reach out to:

Jeriel Tan
PR and Marketing Department

DeepBlue Technology
M: +86 18601755693
E: [email protected]

About DeepBlue Technology
DeepBlue Technology (Shanghai) Ltd was founded in 2014 and is a fast-growing Artificial Intelligence (AI) company with the aim of becoming a world class AI maker platform. Using the guiding principle ofServing humanity through AI, DeepBlue Technology has made breakthroughs in research development and application in various industries, such as Autonomous Driving, Robotics, AI CITY, Biometrics, and AI Retail amongst others. DeepBlue is also a proud recipient of many top AI competitions worldwide, such as PAKDD, IEEE ISI and CVPR.

[email protected]

SOURCE: DeepBlue Technology

Topic:
Company Update
Back to newsroom
Back to Newsroom
Share by: