A self-supervised learning system for category detection by sensory integration
Kouichirou Yamauchi · 1996
Artificial neural network is a useful tool for pattern recognition because the network can realize nonlinear mapping between input and output spaces. This ability is tuned by supervised learning methods such as back-propagation. In the supervised learning methods, desired outputs of the neural network are needed. However, the desired outputs are usually unknown in unpredictable environments. To solve this problem, this paper presents a self-supervised learning system for category detection. This system learns categories of objects and boundaries between them automatically by integrating information from several sensors. We assume that these sensory inputs are always ambiguous patterns which include some noises according to deformation of the objects. After the learning, the system recognizes objects with controlling a priority of each sensor according to the deformation of the sensory input pattern. 1 Introduction Multi-layered neural network is a useful tool for pattern recognition ...