On November 14, the 2018 Text REtrieval Conference (TREC 2018) was held in Gaithersburg, Maryland, USA. The conference announced the results of a total of 7 evaluations in 2018. Among them, three items of the research group of Professor Wu Ji from the Department of Electronic Engineering, Tsinghua University, ranked first, two items ranked second among the six indictors of the two sub-tasks of the TREC 2018 Precision Medicine (PM) Track, ranking first among all participating teams.

Group photo, from left to right: Shot by Chen Xin, Zhou Xuesi, Zhao Gang, Song Jian, Zhao Ling

TREC is a series of text retrieval conferences co-sponsored by the National Institute of Standards and Technology (NIST), among which, TREC Precision Medicine Track is an international authoritative evaluation focusing on retrieval tasks in the field of biomedicine. This year, a total of 24 teams participated in the Scientific Abstracts Task and a total of 21 teams participated in the Clinical Trials Task.

Doctoral student Zhou Xuesi of Grade 2016, master students Song Jian, Chenxin of Grade 2017 and engineer Zhao Gang from the research group of Wu Ji from the Department of Electronic Engineering teamed up to participate in TREC 2018 PM track. Under the guidance of Professor Wu Ji, through introducing medical knowledge base, hierarchical rule optimization and PM literature classifier based on deep neural networks, the system performance has been significantly improved and good results have been achieved.