A novel identification algorithm is proposed to deal with the modelling issues of linear switched systems under the switching modes being unknown.The main contributions of this paper embody the following aspects:(1)fo...A novel identification algorithm is proposed to deal with the modelling issues of linear switched systems under the switching modes being unknown.The main contributions of this paper embody the following aspects:(1)for simplifying the identification problems,the label functions are proposed to replace the switching states of the modes;(2)The recursive least square algorithm is used to identify the system parameters,and the convergence performance of the proposed algorithm is studied.By using convergence analysis,the parameter estimation error can converge to zero point when the input signals satisfy persistent excited conditions and the total time of mode mismatch is limited.Finally,the effectiveness of the proposed method is verified by a simulation example.展开更多
本文提出一种新的基于特征点检测的参数化方法,针对不同面部表情,在三维面部模型存在丢失数据的情况下,改善面部识别准确度。然后采用混合插值方法,对FRGC(Face Recognition Grand Challenge,人脸识别大赛)数据库中的4950幅人脸图像进...本文提出一种新的基于特征点检测的参数化方法,针对不同面部表情,在三维面部模型存在丢失数据的情况下,改善面部识别准确度。然后采用混合插值方法,对FRGC(Face Recognition Grand Challenge,人脸识别大赛)数据库中的4950幅人脸图像进行了人脸特征点实验。Iterative Closest Point定位结果和特征点位置的估计数据证实了该方法的有效性。展开更多
Based on tree-inclusion matching, retrieval may be transformed into matching between the query tree and the integrable-ware label tree. Considering the retrieval specialities of integrable-ware, three theorems of matc...Based on tree-inclusion matching, retrieval may be transformed into matching between the query tree and the integrable-ware label tree. Considering the retrieval specialities of integrable-ware, three theorems of matching are given. On this basis, the inverted-path string algorithm for the integrable-ware label tree query is proposed. This algorithm searches from leaf nodes rather than from root nodes, and considers about the path length and the total number of leaf nodes. It can terminate the failed matching as early as possible and avoid spending too much time on loop comparisons in character string matching. It utilizes the dictionary suffix order to skip much of the impossibility matching path. The experimental results show that this algorithm enhances the recall and the precision of integrable-ware query efficiency while maintaining the searching speed of the integrable-ware.展开更多
Multi-label learning is an effective framework for learning with objects that have multiple semantic labels, and has been successfully applied into many real-world tasks, In contrast with traditional single-label lear...Multi-label learning is an effective framework for learning with objects that have multiple semantic labels, and has been successfully applied into many real-world tasks, In contrast with traditional single-label learning, the cost of la- beling a multi-label example is rather high, thus it becomes an important task to train an effective multi-label learning model with as few labeled examples as possible. Active learning, which actively selects the most valuable data to query their labels, is the most important approach to reduce labeling cost. In this paper, we propose a novel approach MADM for batch mode multi-label active learning. On one hand, MADM exploits representativeness and diversity in both the feature and label space by matching the distribution between labeled and unlabeled data. On the other hand, it tends to query predicted positive instances, which are expected to be more informative than negative ones. Experiments on benchmark datasets demonstrate that the proposed approach can reduce the labeling cost significantly.展开更多
基金supported byNational Natural Science Foundation of China[grant number 61863034].
文摘A novel identification algorithm is proposed to deal with the modelling issues of linear switched systems under the switching modes being unknown.The main contributions of this paper embody the following aspects:(1)for simplifying the identification problems,the label functions are proposed to replace the switching states of the modes;(2)The recursive least square algorithm is used to identify the system parameters,and the convergence performance of the proposed algorithm is studied.By using convergence analysis,the parameter estimation error can converge to zero point when the input signals satisfy persistent excited conditions and the total time of mode mismatch is limited.Finally,the effectiveness of the proposed method is verified by a simulation example.
基金The National High Technology Research and Devel-opment Program of China (863Program)(No2002AA111010)
文摘Based on tree-inclusion matching, retrieval may be transformed into matching between the query tree and the integrable-ware label tree. Considering the retrieval specialities of integrable-ware, three theorems of matching are given. On this basis, the inverted-path string algorithm for the integrable-ware label tree query is proposed. This algorithm searches from leaf nodes rather than from root nodes, and considers about the path length and the total number of leaf nodes. It can terminate the failed matching as early as possible and avoid spending too much time on loop comparisons in character string matching. It utilizes the dictionary suffix order to skip much of the impossibility matching path. The experimental results show that this algorithm enhances the recall and the precision of integrable-ware query efficiency while maintaining the searching speed of the integrable-ware.
文摘Multi-label learning is an effective framework for learning with objects that have multiple semantic labels, and has been successfully applied into many real-world tasks, In contrast with traditional single-label learning, the cost of la- beling a multi-label example is rather high, thus it becomes an important task to train an effective multi-label learning model with as few labeled examples as possible. Active learning, which actively selects the most valuable data to query their labels, is the most important approach to reduce labeling cost. In this paper, we propose a novel approach MADM for batch mode multi-label active learning. On one hand, MADM exploits representativeness and diversity in both the feature and label space by matching the distribution between labeled and unlabeled data. On the other hand, it tends to query predicted positive instances, which are expected to be more informative than negative ones. Experiments on benchmark datasets demonstrate that the proposed approach can reduce the labeling cost significantly.
