Institute of Robotics and Automatic Information System
Tianjin Key Laboratory of Intelligent Robotics
Seminar Series：Advanced Robotics & Artificial Intelligence
报告题目：Research of Human-Scale Brain Simulation for Exascale Supercomputer
报告地点：人工智能学院南楼 329 教室
报告团队：Zhe Sun, Cesar F. Caiafa, Jordi Solé-Casals
Abstract：In the past years, more and more people realized the importance of brain research. Many countries started brain research projects. Even today, many artificial intelligence progresses and deep learning applications show brain-like ability in many areas, but they are still not realistic intelligent systems. On the other hand, the study of brain’s diseases and other medical related topics confront a lot of social pressure. Although brain researches have made great progresses in the past 20 years, we still don’t have well-established theory about how the brain works. In brain research, computational neuroscience plays a very important role, because the process of neural information has a complex spatial-temporal structure. In computational neuroscience, constructing large scale simulation is an important method on understanding and reproducing artificially brain functions. Based on the current flagship supercomputer K computer, our group have succeeded to simulate a neural network with 1.73 neurons and 10.4 trillion synapses, which represents 1% of neural network of the whole human brain. And a simulation consisting of basal ganglia circuits in health and Parkinson’s disease has been developed.
Zhe Sun is currently a research scientist with the Computational Engineering Applications Unit, R&D Group, Head Office for Information Systems and Cybersecurity, National Institute of RIKEN, Japan. He is a kernel researcher of human-scale whole brain simulation on POST-K supercomputer. Zhe Sun received the Master and Ph.D. degree from Yokohama City University, Japan. He joined RIKEN, in 2015, as a Research Support Assistant. He has been a Research Scientist with RIKEN, since 2017. His current research interests include large scale brain simulation, neuromorphic engineering, and high-performance computing system.
Jordi Solé-Casals received the Ph.D. degree with European label in 2000, and the B.Sc. degree in Telecommunications in 1995, both from the Polytechnic University of Catalonia (UPC), Barcelona; and the B.Humanities in 2010 from the Open University of Catalonia (UOC), Barcelona. In 1994 he joined the Engineering Department of the University of Vic - Central University of Catalonia, of which he was Director (2010-2012), and is currently Full Professor and Head of the Data and Signal Processing group. His research interests are in biomedical signal processing (EEG, fMRI, speech, handwritten, biometric applications), neural networks, source separation and independent component analysis.
Cesar F. Caiafa works on the development of machine learning algorithms exploiting tensor decompositions and sparsity with diverse applications ranging from Neuroscience to Astronomy. He received the Ph.D. degree in engineering from the Faculty of Engineering, University of Buenos Aires, Argentina, in 2007. He worked as Research Scientist (2008 – 2010) and Visiting Scientist (2011 – 2018) at Lab. for Advanced Brain Signal Processing, BSI-RIKEN, Wako, Japan. He worked at the Psychology and Brain Sciences Department at the Indiana University, Bloomington, Indiana, USA during 2016 – 2018. He currently holds a permanent position as Independent Researcher with the Argentinean Radioastronomy Institute (IAR) - CONICET and Adjunct Professor with Faculty of Engineering, University of Buenos Aires. He is also Visiting Scientist at the Tensor Learning Unit, at the RIKEN Center for Advanced Intelligence Project, Tokyo, Japan.