Associate Learning Law in a Memristive Neural Network Conference Paper uri icon

abstract

  • In this paper, the Max-Input-Feedback (MIF) algorithm is further studied. It is shown that the choice of the feedback function plays a vital role in improving the performance of the MIF law. By constructing a simple memristive neural network (MNN), trained by the MIF law, to implement the modified Pavlov experiment, a preliminary design criterion of the feedback function is obtained. The effects caused by the parameters of the feedback function on the learning and correcting processes are established. It is indicated that faster learning and correcting speeds can be achieved by choosing a proper feedback function. It is expected that it may provide a guide to the potential applications of the MIF law. 2013 IEEE.

name of conference

  • 2013 Sixth International Conference on Advanced Computational Intelligence (ICACI)

published proceedings

  • 2013 SIXTH INTERNATIONAL CONFERENCE ON ADVANCED COMPUTATIONAL INTELLIGENCE (ICACI)

author list (cited authors)

  • Liu, Y., Huang, H. e., & Huang, T.

citation count

  • 0

complete list of authors

  • Liu, Yujie||Huang, He||Huang, Tingwen

publication date

  • October 2013