Showing papers by "Xiaoou Li published in 2005"

Recurrent neural networks training with stable risk-sensitive Kalman filter algorithm

Abstract: Compared to normal learning algorithms, for example backpropagation, Kalman filter-based algorithm has some better properties, such as faster convergence. In this paper, Kalman filter is modified with a risk-sensitive cost criterion, we call it as risk-sensitive Kalman filter. This new algorithm is applied to train recurrent neural networks for nonlinear system identification. Input-to-state stability is used to prove that the risk-sensitive Kalman filter training is stable. The contributions of this paper are: 1) the risk-sensitive Kalman filter is used for the state-space recurrent neural networks training, 2) the stability of the risk-sensitive Kalman filter is proved.

An Active Rule Base Simulator based on Petri Nets

Abstract: Event-condition-action d e s , in active database systems, should be performed carefully, because iheir firings can produce inconsistent states in database systems. In this paper, an ECA rule base simulator is described, named ECAPNSim, which uses a Conditional Colored Petri Net model to depict ECA rules. It can model ECA rules, simulate their behavior, and perform a static analysis.

Continuous-time recurrent multilayer perceptrons for nonlinear system identification

Abstract: In this paper continuous-time recurrent multilayer perceptrons (RMLP) are proposed to identify nonlinear systems. Using the function approximation theorem for multilayer perceptrons(MLP), we conclude that RMLP can approximate any dynamic system in any degree of accuracy. By means of a Lyapunov-like analysis, a stable learning algorithm for RMLP is determined. The suggested learning algorithm is similar to the well-known backpropagation rule of the multilayer perceptrons but with an additional term which assure the stability of identification error