Machine learning for interactive systems and robots: a brief introduction
Proceedings of the 2nd Workshop on Machine Learning for Interactive Systems: Bridging the Gap Between Perception, Action and Communication
| Hi-index | 0.00 |
In indoor applications, a service robot is required to be able to understand and complete an open-ended set of user tasks. In this case, the designer cannot predict all user tasks, all variants of environment, or what knowledge will be needed in order for the robot to complete one of these tasks. In this talk, I show that open knowledge, i.e., knowledge from open-source resources, is needed and can be employed to meet some challenges from these requirements. We identified some essential research issues and implemented a set of techniques on our OK-KeJia prototype, including multimode NLP, integrated decision-making, and open knowledge searching. Experiments with large test sets (11,615 tasks and 467 desires input by Internet users) showed that open knowledge can be utilized to increase the robot's performance remarkably.