Grid Inc is a Tokyo based technology startup founded in 2009. to create an Artificial Intelligence (AI) platform focused on innovating infrastructure. Grid Inc. was founded on the idea that as the world becomes larger and more advanced, infrastructure must also be evolving and innovating in order to support new information flows and technologies. Deep learning (DL) is making huge waves in the machine learning and computer science community, with a steadily growing impact. In parallel with strides in DL, the field of deep reinforcement learning (DRL) has emerged and led to significant achievements since 2013. In 2016, a DRL-based algorithm beat the world’s (human) “Go” champion, illuminating the potential of reinforcement learning and expanding the scope of its applicability. In his talk, Dr. Sogabe will discuss the background and development of “∞ReNom,” Grid’s deep learning framework for industrial and engineering optimization. While most open source deep learning frameworks focus on pattern recognition and supervised/unsupervised learning application, industrial and engineering data is usually non-graphic and numerically multivariate. Dr. Sogabe will discuss the difficulties of applying deep learning frameworks to this kind of data, and how Grid has overcome this issue using topological graphic representation. Features in ∞ReNom based on topological visualization and analysis enables the creation of a generalized communication interface between humans, machines and society, thus lowering the communication barriers between human operators and machines and allowing easy participation by people from different fields and various backgrounds. Dr. Sogabe will outline Grid’s business vision and research direction for the next five years, and discuss the cooperation opportunities between Japan and U.S.A from a global business point of view. 

SPEAKER

Tomah Sogabe, Co-founder, GRID Inc.

BIO

Tomah Sogabe is currently an associate professor at the University of Electro-communications in Japan. He received his Ph.D. in physics from the Institute for Molecular Science (The Graduate University for Advanced Studies, Japan) in 2007. Dr. Sogabe worked as a postdoc fellow at the Max Planck Institute of Microstructure Physics in Halle, Germany, and at the University of Cambridge, before co-founding Grid Inc. in 2009, where he served as Chief Technology Officer. In 2011, Dr. Sogabe joined the Department of New Energy at the University of Tokyo’s Research Center for Advanced Science and Technology (RCAST), where he worked as an assistant professor and associate professor until 2016. During this period, he served as chief scientist and successfully developed the world’s most efficient quantum dot solar cell and invented the first epitaxial-lift off type quantum dot solar cell. In 2015, Dr. Sogabe founded the artificial intelligence (AI) lab at Grid Inc. and has served as the leading developer of the “∞ReNom” deep learning framework since. Dr. Sogabe is currently leading a team to develop the quantum algorithm based FPGA for implementing deep learning neural networks and the pertinent deep reinforcement learning algorithms. Dr. Sogabe has published more than 40 peer-reviewed papers in international journals including Phys. Rev. Lett., Phys. Rev. B, Prog. Photovolt: Res. Appl., App. Phys. Rev.. He has delivered more than 100 academic talks, international and domestic. His work has been highlighted by Nature: Nanotechnology, Physics (American Physics Society) and cover image letter for Phys. Rev. Lett and etc. Dr. Sogabe is also a member of the Japan Society of Applied Physics (JSAP) and the Japanese Society for Artificial Intelligence. 

AGENDA:

4:15pm: Doors open
4:30pm-5:30pm: Talk and Discussion
5:30pm-6:00pm: Networking

LOCATION:

Philippines Conference Room, 3rd floor Encina Hall, 616 Serra Street, Stanford, CA 94305

PARKING: 

Open parking at Stanford University available starting 4:00pm near Encina Hall unless otherwise marked. Nearest parking garage is Structure 7, below the Graduate School of Business Knight School of Management.

RSVP REQUIRED