which complement the incomplete MBD, and achieve both the shortest design turn-around-time (TAT) and the best system performance. Also in the system operations, the aging performance deteriorations and their variations are difficult to predict in the design process are detected by IoT and AI. The results are not only reported to the system operator but also act as feedback to the next system design, and the system reliability is further improved. This research is in the cross research domain, and the chances that the applied researcher is familiar with all the related layers are slim. The Fellowship researcher will pursue research based-on his/her wide, deep knowledge, rich ideas and systematic thinking regardless of one’s major. This research area is likely to be one of the most important ones for the next generation of Toshiba with its wide variety of power electronics businesses. Moreover, Toshiba is the best environment for such research activity considering the existence of all the related layers of power electronics technologies and applications..Machine learning for collaborative and secure AI
The use of AI-related technologies is widely expected to lead an evolutional change in value creation and to counter the lack of human workforce in manufacturing, logistics, services and many other commercial fields. Specifically, implementing AI into autonomous driving vehicles is considered to be one of the key technologies to significantly improve traffic safety, transportation efficiency and driver/passenger comfort.
Since autonomous driving vehicles have life and death safety implications, AI in this area is required to be reliable in order to prevent unexpected accidents, terror attacks and so on. At the same time, we need to develop a technology to make autonomous AI agents collaboratively work together to achieve common public goals.
Collaborative and secure AI is actively discussed in research communities but most studies are conducted solely in a laboratory setting.
Description of Research
The aim of this programme is to develop and lead new research on collaborative and/or secure AI for autonomous vehicles, which includes problem setting, survey, algorithm development and assessment.
Toshiba has developed an image recognition processor, Visconti(TM). The Toyota Alphard/Vellfire with Visconti(TM)4 has won the JNCAP(*1) award for preventive safety performance(*2). To develop autonomous vehicles with image recognition processors, we need to investigate how to prevent external attacks by adversarial AI etc. and realize secure AI. Moreover, collaborative AI is also an open issue for the efficient autonomous driving system. The method and approach for adversarial robustness and collaborative efficiency between AIs will be necessary for reliable industrial applications including for autonomous vehicles.
Candidates do not necessarily have to consider both collaborative and secure AI in this programme. For instance, research of just secure AI is acceptable.
*1 2018 Japan New Car Assessment Program
*2 Toyota Alphard/Vellfire with Toshiba’s Advanced Image Recognition Processor Wins Japan’s Highest Award for Preventive Safety Performance (Toshiba Electronic Devices & Storage Corporation)
Required Knowledge and Skills
Candidates are required to have a deep knowledge of machine learning (including adversarial learning, deep learning), preferably combined with a knowledge of security. They should have software implementation skills such as C, C++, C#, MATLAB(R) and/or python(TM) for building algorithms and conducting experiments.
 A. Seki, et al., “Patch Based Confidence Prediction for Dense Disparity Map,” British Machine Vision Conference (BMVC), 2016.
 A. Seki, et al., “SGM-Nets: Semi-Global Matching with Neural Networks,” CVPR, 2017.
 R. Katsuki, et al., “Graph Search Based Local Path Planning with Adaptive Node Sampling,” The 29th IEEE Intelligent Vehicles Symposium (IV2018), 2018.
 A. Kawasaki, et al., “Trajectory Prediction of Turning Vehicles Based on Intersection Geometry and Observed Velocities,” The 29th IEEE Intelligent Vehicles Symposium (IV2018), 2018.
 G. Ros, et al., “Training Constrained Deconvolutional Networks for Road Scene Semantic Segmentation,” arXiv:1604.01545, 2016
 T. Watanabe, “Road Scene Understanding Using Image Recognition for Safety and Autonomous Driving,” The 25th International Display Workshops (IDW’18), 2018.