Dazi Li

A sequential method using multiplicative extreme learning machine for epileptic seizure detection

Epilepsy, one of the most common neurological disorders of the human brain, is unpredictable and irregular. There is much difficulty involved in its detection. Here, a sequential processing feature extraction method and a novel multiplicative extreme learning machine are proposed for using in the electroencephalogram (EEG) classification process towards epileptic seizure detection. Firstly, a discrete wavelet transform (DWT) algorithm based on the frequency decomposition is used to obtain the sub-band wavelet coefficients. Secondly, two-dimensional (2D) and three-dimensional (3D) phase space reconstruction (PSR) of the sub-band are calculated to reveal the nonlinear chaos characteristics of signals in the high dimension. Thirdly, and differing from other statistical methods, the singular values are calculated based on the covariance matrix of 2D or 3D phase space as features that reduce the correlation of each dimension and which demonstrate the crucial variance in the original EEG signals. A combination of the proposed sequential methods can extract the significant features of epileptic seizure signals for classification. Finally, a novel multiplicative extreme learning machine (M-ELM) is proposed for using in the classification process. As compared with the normal ELM, support vector machine (SVM) with different kernels and backpropagation (BP) neural networks, the use of M-ELM can further improve the classification accuracy rate of the seizure signals, seizure free signals and healthy signals from the public dataset. Tests of the proposed epilepsy detection approach can achieve the highest 100% detection accuracy with rapid calculation speed.

Reinforcement learning control with adaptive gain for a Saccharomyces cerevisiae fermentation process

A tight and robust yeast fermentation controller is usually difficult to achieve because of the inherent uncertainty, nonlinear, and time-varying characteristics of the yeast fermentation dynamic process. This paper presented an alternative method for yeast fermentation process control by hybrid reinforcement learning algorithm and fuzzy logic. The fuzzy logic was used to adjust the weighting gain of control action adaptively from reinforcement learning. It led to faster tracking and helped to alleviate the overshoot of the controller. The improved multi-step action Q-learning control algorithm was developed and demonstrated through studies on ethanol concentration control of the yeast fermentation process. Experimental results show that the improved multi-step action Q-learning controller has much lower overshoot, faster tracking, shorter transition, and smoother control signal than the advanced PID controller.

Kernel-Based Least Squares Temporal Difference With Gradient Correction

A least squares temporal difference with gradient correction (LS-TDC) algorithm and its kernel-based version kernel-based LS-TDC (KLS-TDC) are proposed as policy evaluation algorithms for reinforcement learning (RL). LS-TDC is derived from the TDC algorithm. Attributed to TDC derived by minimizing the mean-square projected Bellman error, LS-TDC has better convergence performance. The least squares technique is used to omit the size-step tuning of the original TDC and enhance robustness. For KLS-TDC, since the kernel method is used, feature vectors can be selected automatically. The approximate linear dependence analysis is performed to realize kernel sparsification. In addition, a policy iteration strategy motivated by KLS-TDC is constructed to solve control learning problems. The convergence and parameter sensitivities of both LS-TDC and KLS-TDC are tested through on-policy learning, off-policy learning, and control learning problems. Experimental results, as compared with a series of corresponding RL algorithms, demonstrate that both LS-TDC and KLS-TDC have better approximation and convergence performance, higher efficiency for sample usage, smaller burden of parameter tuning, and less sensitivity to parameters.

Improving convolutional neural network using accelerated proximal gradient method for epilepsy diagnosis

The task of epilepsy diagnosing in medicine by classification of electroencephao-graph (EEG) signals is considered. Since an EEG signal has a large number of dimensions as an input sample vector, many previous classification methods have been proposed as hybrid frameworks, which are structurally complex and computationally expensive. In this paper, convolutional neural network (CNN) is used to realize feature extraction and classification simultaneously. The scheme of CNN is adopted to overcome the curse of dimensionality. Meanwhile, the accelerated proximal gradient method is used to increase the training ratio. Experimental results show that the proposed method achieves ideal accuracy of epilepsy diagnosis and converges faster than CNNs based on traditional gradient back propagation.

Adaptive fractional-order total variation image restoration with split Bregman iteration

Alleviating the staircase artifacts for variation method and adjusting the regularization parameters adaptively with the characteristics of different regions are two main issues in image restoration regularization process. An adaptive fractional-order total variation l1 regularization (AFOTV-l1) model is proposed, which is resolved by using split Bregman iteration algorithm (SBI) for image estimation. An improved fractional-order differential kernel mask (IFODKM) with an extended degree of freedom (DOF) is proposed, which can preserve more image details and effectively avoid the staircase artifact. With the SBI algorithm adopted in this paper, fast convergence and small errors are achieved. Moreover, a novel regularization parameters adaptive strategy is given. Experimental results, by using the standard image library (SIL), the lung imaging database consortium and image database resource initiative (LIDC-IDRI), demonstrate that the proposed methods have better approximation, robustness and fast convergence performances for image restoration.

Regularization and feature selection in least squares temporal difference with gradient correction

We focus on the learning prediction problems in reinforcement learning with linear function approximation. In particular, the ℓ1-regularized problems in least-squares temporal difference with gradient correction (LS-TDC) are studied. Since LS-TDC contains gradient correction term, the convergence rate of LS-TDC is higher than that of least-squares temporal difference (LS-TD) algorithm. However, LS-TDC may over-fit to data as LS-TD does when the number of features is larger than that of samples. Thus, the regularization and feature selection of LS-TDC are studied. It is well known that ℓ1-regularization can produce sparse solutions and often serves as an automatic feature selection method in value function approximation. The ℓ1-regularized problem in LS-TDC adds a penalty term into the fixed-point function, but this augment function cannot be solved analytically. We turn to build the optimal solution incrementally by using an algorithm similar to Least Angle Regression (LARS) algorithm and LARS-TD algorithm. By using LARS algorithm, an ℓ1-regularized version of LS-TDC named LARS-TDC is proposed. Experiment results show that LARS-TDC is an effective method to solve the ℓ1-regularized problem.

Least-squares temporal difference learning with eligibility traces based on regularized extreme learning machine

The task of learning the value function under a fixed policy in continuous Markov decision processes (MDPs) is considered. Although ELM has fast learning speed and can avoid tuning issues of traditional artificial neural network (ANN), the randomness of the ELM parameters would result in fluctuating performance. In this paper, a least-squares temporal difference algorithm with eligibility traces based on regularized extreme learning machine (RELM-LSTD(X)) is proposed to overcome these problems caused by ELM in Reinforcement Learning problem. The proposed algorithm combined the LSTD(X) algorithm with RELM. The RELM is used to approximate value functions. Furthermore, the eligibility trace term is introduced to increase data efficiency. In experiments, the performances of the proposed algorithm are demonstrated and compared with those of LSTD and ELM-LSTD. Experiment results show that the proposed algorithm can achieve a more stable and better performance in approximating the value function under a fixed policy.

Quasi-Linear Extreme Learning Machine Model Based Nonlinear System Identification

A regression algorithm of quasi-linear model with extreme learning machine (QL-ELM) and its applications for nonlinear system identification are presented. The distinctive feature of the proposed method is that the Quasi-linear model is constructed as a linear ARX model with a complicate nonlinear coefficient. It not only has various linearity properties but also shows some good approximation ability. The complicated coefficients are separated into two parts. The linear part is determined by recursive least square, while the nonlinear part is identified through extreme learning machine. The whole methodology is presented in detail. The effectiveness and accuracy of the proposed method is extensively verified in two nonlinear system identification, including a chemical continuously stirred tank reactor (CSTR) process.

Application of hybrid prediction model for dry point soft sensing of aviation kerosene

In this paper, an approach to build soft sensing prediction model for the dry point of aviation kerosene in a complex non-linear hydrocracking unit is proposed. PLS, RBFN and PLS-RBFN soft sensor models are first established for dry point of aviation kerosene. On this basis, the three sub-models are combined into a hybrid soft sensing prediction model through principal components regression. The performance of the soft sensor models based on hybrid soft sensing prediction model is compared with that of PLS, RBFN and PLS-RBFN respectively. Numerical results showed that soft sensor model by hybrid prediction model has better forecast accuracy and model stability than the other three methods, and can be well adaptive to the changing working conditions.

On-line Optimization of Fuzzy-Immune PID for PEMFC Temperature Control Based on RBF Neural Network

Proton Exchange Membrane Fuel Cell (PEMFC) temperature exists complex nonlinearity and is deeply disturbed by load change. Considering the characteristics of PEMFC temperature control, an improved fuzzy-immune PID algorithm is derived based on the immune feedback regulating law. Compared with general fuzzy-immune PID algorithm, radial basis function (RBF) neural network is introduced to the on-line optimization work of fuzzy-immune PID parameters, which optimizes the PID parameters on-line. Simulation results show that the proposed method in this study achieves good performance in temperature control and is useful for wide application of PEMFC.