Yulai Zhang

Argument Ranking with Categoriser Function

Recently, ranking-based semantics is proposed to rank-order arguments from the most acceptable to the weakest one(s), which provides a graded assessment to arguments. In general, the ranking on arguments is derived from the strength values of the arguments. Categoriser function is a common approach that assigns a strength value to a tree of arguments. When it encounters an argument system with cycles, then the categoriser strength is the solution of the non-linear equations. However, there is no detail about the existence and uniqueness of the solution, and how to find the solution (if exists). In this paper, we will cope with these issues via fixed point technique. In addition, we define the categoriser-based ranking semantics in light of categoriser strength, and investigate some general properties of it. Finally, the semantics is shown to satisfy some of the axioms that a ranking-based semantics should satisfy.

Short term power load prediction with knowledge transfer

A novel transfer learning method is proposed in this paper to solve the power load forecast problems in the smart grid. Prediction errors of the target tasks can be greatly reduced by utilizing the knowledge transferred from the source tasks. In this work, a source task selection algorithm is developed and the transfer learning model based on Gaussian process is constructed. Negative knowledge transfers are avoided compared with the previous works, and therefore the prediction accuracies are greatly improved. In addition, a fast inference algorithm is developed to accelerate the prediction steps. The results of the experiments with real world data are illustrated.

A New Recursive Kernel Regression Algorithm and Its Application in Ultra-short Time Power Load Forecasting

Abstract A new kernel regression algorithm is introduced in this work. The algorithm combines the kernel method and the recursive least squares method, and it is especially useful for on-line time series forecasting and non-linear systems. Under the interruption of colored noises, this algorithm performed better than existing algorithms of the same kind. We validated the algorithms accuracy and time cost using both numerical simulations and real-world experimental data.

Power Load Forecasting based on Multi-task Gaussian Process

Abstract The power load data from nearby cities are significantly correlated because they share the same hidden variables as well as the underlying noises. A multi-task Gaussian process method for non-stationary time series prediction is introduced and applied to the power load forecasting problem in this paper. The prediction accuracies are effectively improved due to the additional information provided by the related data sets. A novel algorithm for prediction is developed to reduce the computational complexity of the multi-task Gaussian Process method. The algorithms prediction precision and efficiency are validated by a real world short-term power load data sets.

An Entropy Based Method for Causal Discovery in Linear Acyclic Model

A new algorithm for causal discovery in linear acyclic graphic model is proposed in this paper. The algorithm measures the entropy of observed data sequences by estimating the parameters of its approximate distribution to a generalized Gaussian family. Causal ordering can be discovered by an entropy base method. Compared with previous method, the sample complexity of the proposed algorithm is much lower, which means the causal relationship can be correctly discovered by a smaller number of samples. An explicit requirement of data sequences for correct causal inference in linear acyclic graphic model is discussed. Experiment results for both artificial data and real-world data are presented.

Recursive prediction algorithm for non-stationary Gaussian Process

The original exact inference algorithm of the GP model runs very slow.We developed a recursive inference algorithm as an improvement.The new algorithm can obtain the same result in a shorter time.It works well for the real-time online prediction problems. Gaussian Process is a theoretically rigorous model for prediction problems. One of the deficiencies of this model is that its original exact inference algorithm is computationally intractable. Therefore, its applications are limited in the field of real-time online predictions. In this paper, a recursive prediction algorithm based on the Gaussian Process model is proposed. In recursive algorithms, the computational time of the next step can be greatly reduced by utilizing the intermediate results of the current step. The proposed recursive algorithm accelerates the prediction and avoids the loss of accuracy at the same time. Experiments are done on an ultra-short term electric load data set and the results are demonstrated to show the accuracy and efficiency of the new algorithm.

Inferring causal directions from uncertain data

Abstract Causal knowledge discovery is an essential task in many disciplines. Inferring the knowledge of causal directions from the measurement data of two correlated variables is one of the most basic but non-trivial problems in the research of causal discovery. Most of the existing methods assume that at least one of the variables is strictly measured. In practice, uncertain data with observation error is widely exists and is unavoidable for both the cause and the effect. Correct causal relationships will be blurred by such noise. A causal direction inference method based on the errors-in-variables (EIV) model is proposed in this work. All variables are assumed to be measured with observation errors in the errors-in-variables models. Causal directions will be inferred by computing the correlation coefficients between the regression model functions and the probability density functions on both of the possible causal directions. Experiments are done on artificial data sets and the real world data sets to illustrate the performance of the proposed method.

A New Recursive Identification Method for Weighted Criterion

Abstract In time-series system modeling traditional criterions only consider current estimation error while the past and global prediction errors are usually overlooked. By integrating both the estimation error and the difference between neighbor prediction errors, a novel weighted identification criterion was presented. Based on this criterion the alternations of the two-step algorithm is constructed through separating the system parameters estimation and noise parameters estimation for the system disturbed by colored noise, which could result in oscillation and instability. An extended one-step recursive algorithm for the weighted identification criterion is introduced in this paper. For the input-output system disturbed by colored noise, the prediction gradients and the gradient of the pseudo linear regression vector are given. The gradient iterative algorithm and the direct adaptive method (DAM), the new one-step recursive algorithm are proposed by a series of estimation process optimizations. Finally, a simulation example is conducted to demonstrate the efficiency of this new method.

Social Welfare Semantics for Value-Based Argumentation Framework

Logically intuitive properties of argument acceptance criteria have been paid great attention in the argumentation frameworks, however, which do not take into account the agents. Recently research has begun on evaluating argument acceptance criteria considering preferences of audiences, a kind of representative agents. In this paper we take a step towards applying social choice theory into value-based argumentation framework (VAF), and propose social welfare semantics, an extension of Dung’s semantics incorporating social preference. Then the social welfare semantics and its relationship with the semantics of VAF are analyzed, and some non-trivial properties are proved based on Pareto efficiency.

Self-tuning filtering for multi-sensor data fusion based on forget factor algorithms

The existing algorithms of data fusion will face the problem of data saturation when interrupted by noises with large variance. Multi-sensor data fusion, which can address this issue, is examined in this paper. The forget factor (FF) method was introduced into the data fusion algorithm to avoid the data saturation phenomenon. A proof for the sequence equivalence theory was given, which showed that two data sequences with different orders can be equivalent to a single sequence whose order is the same as the higher one. In the simulations, an optimal fusion method was used to show the advantages of the algorithm for parameter estimation under large-variance noises.