How to assess the performance of machine learning algorithms is a problem of increasing interest and urgency as the data mining application of myriad algorithms grows. The standard...
Selecting the optimal kernel is an important and difficult challenge in applying kernel methods to pattern recognition. To address this challenge, multiple kernel learning (MKL) ...
In this paper, we apply an evolutionary algorithm to learning behavior on a novel, interesting task to explore the general issue of learning e ective behaviors in a complex enviro...
Singularity is the special feature of fingerprints for identification and classification. Since the performance of singularity extraction depends on the quality of fingerprint imag...
Abstract— Continuous action sets are used in many reinforcement learning (RL) applications in robot control since the control input is continuous. However, discrete action sets a...
Akihiko Yamaguchi, Jun Takamatsu, Tsukasa Ogasawar...