【Turing Seminar Term No. Fourteen】Applying of Schematic Learning in Medicinal Discovery


Report time: 3:00 p.m. January 8, 2021

Report form: Tencent conference

Conference ID: 980 355 875

Report Topic:  Applying of Schematic Learning in Medicinal Discovery

Reported by: Zhang Ming (Professor in Beijing University)

About the Content:  Many basic assignments in the sector of new pharmaceutical discovery are closely associated with schematic learning algorithm and graphic neutral network study. The report starts by specifying molecular diagram assignment in new pharmaceutical discovery sector by presenting the first GraphAF, the first model generated based on autoregression flow that has grabbed the best effect in generation of multiple molecular diagrams. It follows by designing a model G2Gs that does not hinge on any chemical domain knowledge for the assignment of molecular synthesis route design. Its performance resembles that of template-based methods, but does not rely on domain knowledge like those template methods. This model has strong interpretability and scalability. This work was completed by Shi Chence, a student of the first Turing Class of Peking University, as the first author in his undergraduate graduation thesis. Relevant results were published in the top conferences namely ICLR2020 and ICML2020 in machine learning and medical care. It attracted extensive attention from scholars in the field.

About the Lecturer:  Professor Zhang ming, from the School of Electronics Engineering and Computer Science of Peking University, doctoral supervisor, member of Computer Course Teaching Committee of the Ministry of Education, and sole Chinese director of the ACM Education Committee, she was enrolled to Peking University in 1984 and received bachelor's, master's and doctorate degrees in here. Her research orientation is text mining, machine learnin and so on.She currently presides over National Science and Technology Ministry Science and Technology Innovation 2030-’New Generation Artificial Intelligence’, a major project sub-topic, National Natural Science Foundation of China and other projects, co-published more than 200 scientific research academic papers, and has been cited more than 9,800 times in Google Academic with an H factor being 39. She won over the best paper award at ICML 2014, a top conference on machine learning, and was nominated for best paper in WWW 2016, a top conference on network information processing. Her network representation model LINE published in WWW 2015 has been cited for over 2,800 time. She won the first place in highly-quoted thesis in 2015-2019 in WWW conference. Data Structure and Algorithm she lectured was selected national and Beijing-level quality courses, state-level courses with fancy resources, national quality online courses and state first-class undergraduate courses.