Abstract. As deep learning is increasingly used in safety-critical systems including self-driving cars, medical diagnosis and malware detection, its reliability has become a big concern. In this talk, we first outline some reliability problems in deep learning systems, and then introduce some automated testing techniques we developed to ensure the reliability of deep learning systems.

Biography. Jianjun Zhao received a B.E. degree in Computer Science from Tsinghua University (China) in 1987, and a Ph.D. degree in Computer Science from Kyushu University (Japan) in 1997. After receiving his Ph.D., he joined the Department of Computer Science and Engineering, Fukuoka Institute of Technology (Japan) as an Assistant Professor in 1997, and then was promoted to be an Associate Professor in 2000. Since November 2005, he had been with the School of Software, and then the Department of Computer Science and Engineering, Shanghai Jiao Tong University (China) as a full professor. From April, 2016, he joined the Department of Advanced Information Technology, Kyushu University as a full professor. He is also a visiting scientist in the MIT Laboratory for Computer Science (USA) from April 2002 to March 2003, a guest professor of National Institute of Informatics (Japan) from April 2012 to March 2017, and a visiting professor of Shanghai Jiao Tong University (China) since 2016. His main research interest is software engineering and programming language, in particular reliable deep learning systems, program analysis and verification, automatic programming, software testing, debugging, and programming environments. He has published more than 80 research papers in a number of conferences proceedings, including ICSE, FSE, PLDI, ECOOP, ASE, ISSTA, and ICSM. He also served as a PC member for various international conferences including ICSE, FSE, ASE, ECOOP, ICSM, ISSRE, and WICSA.

Abstract. Social coding is widely used in software development. Various users are suggesting contributions to improve the source code of the social coding project. Not only the proposal itself but also the contributors' reputation influence the decision on approval/rejection of these contributions. In this talk, we show the index of contribution excluding prejudice based on contributor's reputation, and show how to measure index by learning behavior of contributor. In addition, I will present some research topics on Japanese groupware.

Biography. Yoshinari Nomura received the B.E. degree, the M.E. degree, and the Ph.D. degree in computer science, all from the Kyushu University, Fukuoka, Japan. He had been a research associate of Information Science and Electrical Engineering at Kyushu University. He is currently an Associate Professor of Faculty of Engineering at Okayama University. His interests include operating system, computer network, and groupware systems.

Abstract. In modern smart world, we and our society demand advanced information systems that can give us appropriate and timely help or assistance in our works and everyday life. However, traditional information systems are not such systems because they are passive and reactive, and do not cooperate with human and other systems well. How to realize advanced information systems is challenging research topics.

Biography. Dr. Yuichi Goto is an associate professor of computer science at Graduate School of Science and Engineering, Saitama University in Japan. He received the degree of Bachelor of Engineering in computer science, the degree of Master of Engineering in computer science, and the degree of Doctor of Engineering in computer science from Saitama University in 2001, 2003, and 2005, respectively. His current research interests include relevant reasoning and its applications, automated theorem finding, epistemic programming, anticipatory reasoning reacting systems, and Web services. He is a member of ACM, IEEE-CS, Information Processing Society of Japan (IPSJ), Institute of Electronics, Information and Communication Engineers (IEICE), and Japanese Society of Artificial Intelligence (JSAI).

Abstract. The problem of automated theorem finding is ``What properties can be identified to permit an automated reasoning program to find new and interesting theorems, as opposed to proving conjectured theorems", which is one of 33 basic research problems in automated reasoning originally proposed by Wos in 1988. It is still an open problem. The talk will show various approaches of automated theorem finding, present the state of the art of automated theorem finding, and discuss the problems and difficulties of current works. Current works are all based on the approaches of automated theorem proving and use the existing systems of automated theorem proving or model checking accidentally to find theorems in a certain field. As a new direction of general solution to the problem, the talk will also introduce our approach which is called the reasoning method based on strong relevant logic, and provide an overview of our works.

Biography. Dr. Hongbiao Gao is a lecturer in the school of control and computer engineering at the North China Electric Power University, Beijing, China. He received his Bachelor Degree in engineering from Northeastern University (China) in 2009, and received his Master Degree and Ph.D. in engineering from Saitama University (Japan) in 2012 and 2015, respectively. He worked as an assistant professor in the department of information and computer sciences of Saitama University from 2015 to 2018. His research interests are automated reasoning and automated theorem finding. On the research of automated theorem finding, He obtained research funding supported by Japan Society for the Promotion of Science from 2015 to 2017.