Deng Haojiang

A dynamic web services selection based on decomposition of global QoS constraints

The objective of web services selection is to maximize an application-specific utility function under the end-to-end QoS constraints. Existing methods can be divided into two types: local selection method and global selection method. The local selection method is simple and efficient, but can not meet the end-to-end QoS constraints. The global selection method can satisfy the global QoS constraints, but at the price of higher computational time, and it is not suitable for the dynamic environment. To address this issue, this paper proposes an algorithm to combine global QoS constraints with local selection, which first splits the global QoS constraints into local constraints using heuristic method, and then uses local selection to find the optimal solution under the local constraints. The experimental results show that our approach improves the performance by reducing the computation time greatly while achieving close-to-optimal results.

Priority ordered BP neural network and the application for speaker identification

The backpropagation neural network (BPNN) has been researched and applied to solve the problem that the training time of the backpropagation network can be excessive, so the structure and training algorithm of priority ordered BP neural networks are proposed. The neurons of its output layer have priority ordered interconnections, during the training course, the training data tails off gradually, so the algorithm may converge rapidly because of the decrease of the complexity of performance function. Compared with the conventional BPNN, the total iterative epochs of priority ordered BPNN are far lower and the performance function can converge more rapidly in a text-independent speaker identification task.

A Data-Intensive Multimedia Service Invoking Method Based on SOAP Extending

Nowadays, most of multimedia information processing application systems are still realized in the monolithic and stove-pipe manner. This design pattern makes building and maintaining such multimedia systems intractable, costly and time-consumed. As a new style software design model, Service Oriented Architecture promises a more agile solution for developing large-scale Internet-based applications and has been widely accepted by the academic community and industries. However, the current service invoking method: SOAP does not apply to the multimedia area directly. This paper intends to solve this problem. Aiming at it, this paper proposes a new remote object access protocol based on SOAP which defines the correlative message format and the interaction model between the service and its client. According to the proposed approach, a service-composition-based prototype multimedia information processing system is constructed. Through the experiment, we prove that the performance of our method is excellent compared with other methods.

Combination of likelihood scores using linear and SVM approaches for text-independent speaker verification

In this paper, the text-independent speaker recognition system based on the adapted GMMs was established, and the speaker-independent background model and speaker-dependent models of cohort speaker sets were used to normalize the likelihood score. The approaches to combine likelihood scores using linear and SVM (support vector machine) method in score domain was proposed. The speaker verification experiments over telephone channels showed that based on the likelihood ratio of adapted GMMs system, combination of likelihood scores can improve the verification performance of baseline system using universal background model (UBM). Specially, the approach of score combination using SVM achieved the best performance.