Papers
49
Total Citations
641
H-Index
16
About
Komei Sugiura is a prominent researcher at the intersection of robotics, natural language processing, and human-robot interaction, with a focus on enabling domestic and service robots to understand and respond to human communication in real-world environments. His work has significantly advanced how robots interpret natural language instructions, particularly for everyday tasks like fetching and manipulating objects. A landmark contribution is his research on GAN-based multimodal language understanding, allowing robots to parse ambiguous fetching instructions by jointly reasoning over language and visual input — work that has garnered over 40 citations. Sugiura has also shaped the field of cloud robotics through the Rospeex platform, which democratizes spoken dialogue capabilities for robot developers, and through pioneering more natural, dialogue-oriented robot speech synthesis. His early work on grounded language acquisition — teaching robots to learn object-manipulation verbs from demonstration — laid important conceptual foundations, while more recent contributions explore explainable deep reinforcement learning using attention mechanisms. His involvement with RoboCup@Home competitions (83 citations) reflects a commitment to benchmarking real-world robot performance. Across more than a decade of research, Sugiura has consistently pushed toward robots that communicate naturally and reliably with everyday users.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Rospeex: A cloud robotics platform for human-robot spoken dialogues37 citations · 2015
- 4
- 5Learning object-manipulation verbs for human-robot communication27 citations · 2007
- 6
- 7Learning Novel Objects for Extended Mobile Manipulation20 citations · 2011
- 8
- 9
- 10A cloud robotics approach towards dialogue-oriented robot speech18 citations · 2015