Michiaki Tatsubori  Michiaki Tatsubori photo         

contact information

IBM Research AI
Tokyo, Japan
  

links

Professional Associations

Professional Associations:  ACM

more information

More information:  CV

profile


Michiaki (Mich) Tatsubori leads the Embodied Learning group of the AI department at IBM Research - Tokyo. Just after receiving a PhD degree in Engineering from University of Tsukuba, Japan, he joined the lab in 2002, to work on Web technologies such as web services, server runtimes, and dynamic scripting languages. In 2011, he dived into the area of smarter cities, starting from resilience engineering, open data, machine learning, and those applications for emerging counties.

In 2013, as a new research lab opened in Africa, he decided to move to the lab in Nairobi, Kenya, with his family and engaged in Human Mobility projects with other researchers gathering from world-wide, to enable data-supported transportation solutions leveraging mobile/frugal infrastructure and cutting-edge intelligent transportation technologies. He had lead research technologies-based solution development such as in-car smartphone-based traffic/road analysis.

One of interesting projects embedded smartphone-based tracking devices in garbage collection vehicles. While the tracking itself helped a lot for the Nairobi county government to manage their fleet locations, motion sensors in the smartphones gave more interesting information about the town. By inferring the driving behaviors and road conditions through machine learning on the sensor data, we could obtain a map of road distress in the city, individual driving quality with road condition considered, and even dynamic information of moving traffic obstacles on the road such as human-drawn carts and animals!

He came back to Tokyo in 2017 to enjoy Rugby World Cup and Olympics, and since then has been engaged with embedded AI and its applications such as robotics, visions, planning, and conversations. For example, their robotics solutions combine neural network-based machine learning components, such as object detection and reinforcement-learning, and traditional symbolic planners, leveraging Semantic Web technologies. It allows collaborative robots to learn and plan for effectively solving complicated problems in practice. He is a certified referee of Mini Rugby in Kanagawa, Japan.