Abstract. One of the most difficult problems in multiagent systems involves representing knowledge and beliefs of agents in dynamic environments. New perceptions modify an agent’...
Many image processing and computer vision applications have difficulty dealing with a nonstatic background such as water waves, but this particular dynamic scene actually contains...
— In this paper, we present a novel approach to controlling a robotic system online from scratch based on the reinforcement learning principle. In contrast to other approaches, o...
Clustering aims to find useful hidden structures in data. In this paper we present a new clustering algorithm that builds upon the consistency method (Zhou, et.al., 2003), a semi-...
Abstract. We consider a new discriminative learning approach to sequence labeling based on the statistical concept of the Z-score. Given a training set of pairs of hidden-observed ...