Shaobo Li

An Ensemble Deep Convolutional Neural Network Model with Improved D-S Evidence Fusion for Bearing Fault Diagnosis

Intelligent machine health monitoring and fault diagnosis are becoming increasingly important for modern manufacturing industries. Current fault diagnosis approaches mostly depend on expert-designed features for building prediction models. In this paper, we proposed IDSCNN, a novel bearing fault diagnosis algorithm based on ensemble deep convolutional neural networks and an improved Dempster–Shafer theory based evidence fusion. The convolutional neural networks take the root mean square (RMS) maps from the FFT (Fast Fourier Transformation) features of the vibration signals from two sensors as inputs. The improved D-S evidence theory is implemented via distance matrix from evidences and modified Gini Index. Extensive evaluations of the IDSCNN on the Case Western Reserve Dataset showed that our IDSCNN algorithm can achieve better fault diagnosis performance than existing machine learning methods by fusing complementary or conflicting evidences from different models and sensors and adapting to different load conditions.

Reputation-based multi-dimensional trust model in cloud manufacturing service platform

In previous researches of cloud manufacturing services, trust evaluation models were not flexible enough and subjective preferences of agents were not reflected either. For this issue, a five-dimensional trust evaluation system is built, and a reputation-based trust evaluation algorithm (RBTE) is also proposed. The direct trust rank (DT), friend reputation (RF) and platform reputation (RP) of the agent could be integrated into the trust rank (TR) with the weight vector through this algorithm. Combined with similarities of trust evaluation process, a trust network building approach is put forward in this research. The trust rank update algorithm also achieved dynamic rewards or punishments over time. Experimental results has shown good executing efficiency, and reflected the subjective preference of the transaction agent to a certain extent.

Multi-objective evolutionary algorithm based on decision space partition and its application in hybrid power system optimisation

The distribution of individuals in a population significantly influences convergence to global optimal solutions. However, determining how to maximise decision space information, which benefits convergence, is disregarded. This paper proposes a type of multi-objective evolutionary algorithm based on decision space partition (DSPEA), and designs the sphere initialisation strategies, initialisation method of individuals in each sphere, updating approach for the centroid, radius, and individuals of a hypersphere, and information sharing mechanism among spheres. The decision space in the DSPEA framework is explicitly divided into several hyperspheres. The non-dominated sorting genetic algorithm II is employed to implement each evolution of each hypersphere. An improvement approach related to the information sharing of the spheres is used to produce the future motions of the spheres by adopting particle swarm optimisation. Twelve problems were used to test the performance of DSPEA, and extensive experimental results show that DSPEA performs better than six state-of-the-art multi-objective evolutionary algorithms. Finally, DSPEA is used to optimise a hybrid power system. The results of the simulation optimisation tests on the parameters of the control strategy and the drive system for hybrid electric vehicles demonstrate that the proposed approach can obtain a set of improved solutions with low fuel consumption and pollutant emission.

Patent Keyword Extraction Algorithm Based on Distributed Representation for Patent Classification

Many text mining tasks such as text retrieval, text summarization, and text comparisons depend on the extraction of representative keywords from the main text. Most existing keyword extraction algorithms are based on discrete bag-of-words type of word representation of the text. In this paper, we propose a patent keyword extraction algorithm (PKEA) based on the distributed Skip-gram model for patent classification. We also develop a set of quantitative performance measures for keyword extraction evaluation based on information gain and cross-validation, based on Support Vector Machine (SVM) classification, which are valuable when human-annotated keywords are not available. We used a standard benchmark dataset and a homemade patent dataset to evaluate the performance of PKEA. Our patent dataset includes 2500 patents from five distinct technological fields related to autonomous cars (GPS systems, lidar systems, object recognition systems, radar systems, and vehicle control systems). We compared our method with Frequency, Term Frequency-Inverse Document Frequency (TF-IDF), TextRank and Rapid Automatic Keyword Extraction (RAKE). The experimental results show that our proposed algorithm provides a promising way to extract keywords from patent texts for patent classification.

Heterogeneous System Integration Based on Service Component

For problem of heterogeneous system integration, Service-oriented component of integration model is proposed and the integration framework of component-based is constructed in this paper,which is based on research service component and integrated structure of SOA. The structure of service components is Designed,and analysis its working principle. WEB service and description criterion of registered for system is defined. Paper further study the key technical of heterogeneous information integration and mapping model. Finaly, with examples no heterogeneous systems integrate framework of service component is designed and implemented.

Studies on Fast Pareto Genetic Algorithm Based on Fast Fitness Identification and External Population Updating Scheme

This paper investigates fast Pareto genetic algorithm based on fast fitness identification and external population updating scheme (FPGA) for searching Pareto-optimal set, which is based on a new approach of fast fitness identification algorithm for individual and a clustering on the basis of external population updating scheme to maintain population diversity and even distribution of Pareto solutions. Experiments on a set of multi-objective 0/1 knapsack optimization problems shows that FPGA can obtain high-quality, well distributed nondominated Pareto solutions with less computational efforts compared to other state-of art algorithms, and FPGA in convergence speed outperforms the representative SPEA.

Convolutional Neural Network-Based Embarrassing Situation Detection under Camera for Social Robot in Smart Homes

Recent research has shown that the ubiquitous use of cameras and voice monitoring equipment in a home environment can raise privacy concerns and affect human mental health. This can be a major obstacle to the deployment of smart home systems for elderly or disabled care. This study uses a social robot to detect embarrassing situations. Firstly, we designed an improved neural network structure based on the You Only Look Once (YOLO) model to obtain feature information. By focusing on reducing area redundancy and computation time, we proposed a bounding-box merging algorithm based on region proposal networks (B-RPN), to merge the areas that have similar features and determine the borders of the bounding box. Thereafter, we designed a feature extraction algorithm based on our improved YOLO and B-RPN, called F-YOLO, for our training datasets, and then proposed a real-time object detection algorithm based on F-YOLO (RODA-FY). We implemented RODA-FY and compared models on our MAT social robot. Secondly, we considered six types of situations in smart homes, and developed training and validation datasets, containing 2580 and 360 images, respectively. Meanwhile, we designed three types of experiments with four types of test datasets composed of 960 sample images. Thirdly, we analyzed how a different number of training iterations affects our prediction estimation, and then we explored the relationship between recognition accuracy and learning rates. Our results show that our proposed privacy detection system can recognize designed situations in the smart home with an acceptable recognition accuracy of 94.48%. Finally, we compared the results among RODA-FY, Inception V3, and YOLO, which indicate that our proposed RODA-FY outperforms the other comparison models in recognition accuracy.

Inferring phase diagrams from X-ray data with background signals using graph segmentation

ABSTRACT Automated composition-structure-processing phase diagram creation is critical for high-throughput experimental material studies. In particular, diffractogram datasets with large background signals are especially difficult to identify the phase regions. In this work, we proposed a novel graph segmentation algorithm from computer vision to solve the phase diagram prediction problem from X-ray diffraction data with large background signals. We introduced a novel background subtraction algorithm with graph-based clustering/segmentation to build the BGPhase algorithm. Experiments on three datasets with the Al–Cu–Mo material family showed that our phase attribution algorithm can achieve high prediction accuracy ranging from 88.6 to 94.8% or with MCC scores ranging from 0.715 to 0.890. The algorithm can be accessed online at http://mleg.cse.sc.edu/bgphase.

A Research for Evolutionary Process Based on TRIZ Evolution Theory

Technology Evolutionary Process have own rule and model,it can be forcasted. How to predict the future technological development and quickly develop next-generation products has become a powerful weapon of market competition. TRIZ EvolutionTheory is one of the most advantages and vitality in almost product technology prediction theory. This article summarized the evolution mode and evolution route based on in-depth study of TRIZ evolution theory ,and researched correlation.At last this article introduced how to use the basic principles of evolutionary theory to solve practical problems of the process approach. And to farming with the plow as an example a typical product design, to have been verified.

DeepPatent: patent classification with convolutional neural networks and word embedding

Patent classification is an essential task in patent information management and patent knowledge mining. However, this task is still largely done manually due to the unsatisfactory performance of current algorithms. Recently, deep learning methods such as convolutional neural networks (CNN) have led to great progress in image processing, voice recognition, and speech recognition, which has yet to be applied to patent classification. We proposed DeepPatent, a deep learning algorithm for patent classification based on CNN and word vector embedding. We evaluated the algorithm on the standard patent classification benchmark dataset CLEF-IP and compared it with other algorithms in the CLEF-IP competition. Experiments showed that DeepPatent with automatic feature extraction achieved a classification precision of 83.98%, which outperformed all the existing algorithms that used the same information for training. Its performance is better than the state-of-art patent classifier with a precision of 83.50%, whose performance is, however, based on 4000 characters from the description section and a lot of feature engineering while DeepPatent only used the title and abstract information. DeepPatent is further tested on USPTO-2M, a patent classification benchmark data set that we contributed with 2,000,147 records after data cleaning of 2,679,443 USA raw utility patent documents in 637 categories at the subclass level. Our algorithms achieved a precision of 73.88%.