Home Robots, Learn by Themselves

摘要

To build an intelligent robot, we must develop an autonomous mental development system that incrementally and speedily learns from humans, its environments, and electronic data. This paper presents an ultra-fast, multimodal, and online incremental transfer learning method using the STAR-SOINN. We conducted two experiments to evaluate our method. The results suggest that recognition accuracy is higher than the system that simply adds modalities. The proposed method can work very quickly (approximately 1.5 [s] to learn one object, and 30 [ms] for a single estimation). We implemented this method on an actual robot that could estimate attributes of "unknown" objects by transferring attribute information of known objects. We believe this method can become a base technology for future robots. SOINN is an unsupervised online-learning method capable of incremental learning. By approximating the distribution of input data and the number of classes, a self-organized network is formed. SOINN offers the following advantages: network formation is not required to be predetermined beforehand, high robustness to noise, and reduced computational cost. In the near future, a SOINN device will accompany an individual from birth; this will allow the agent to share personal histories with its owner. In this occasion, a person's SOINN will know "everything" about its owner, lending assistance at any time and place throughout one's lifetime. Besides having a personal SOINN, an individual can install this self-enhanced agent into human-made products - making use of learned preferences to make the system more efficient. If deemed non-confidential, an individual's SOINN could also autonomously communicate another SOINN to share information.