H. Kashioka

About

H. Kashioka is a leading researcher in robot language acquisition and imitation learning, whose work bridges the gap between human communication and autonomous robotic action. His key research areas include active learning for human-robot interaction, probabilistic motion generation, and object manipulation. Kashioka’s major contribution is the development of a confidence measure function that enables robots to actively learn when to ask clarifying questions, thereby reducing the risk of accidents from ambiguous commands in dialogue. This work, published in 2010, has garnered 19 citations and is foundational for safe, adaptive robot behavior. He also pioneered reference-point-dependent trajectory hidden Markov models (HMMs) for imitation learning, allowing robots to generate complex motions—such as rotating or placing objects—by learning from demonstration. This 2011 paper, with 4 citations, advances probabilistic modeling in robotics. Kashioka’s research is notable for its practical impact on creating more intuitive and reliable robotic assistants, making him a key figure in the field of interactive robotics and machine learning.

Research Focus

Key Achievements

Top Papers

1. 1
   Active learning of confidence measure function in robot language acquisition framework

   19 citations · 2010
2. 2
   Motion generation by reference-point-dependent trajectory HMMs

   4 citations · 2011

Key Collaborators

Contact & Links