The number of word embedding models is growing every year. Most of them are based on the co-occurrence information of words and their contexts. However, it is still an open question what is the best definition of cont...
详细信息
Concept-cognitive learning (CCL) is a hot topic in recent years, and it has attracted much attention from the communities of formal concept analysis, granular computing and cognitive computing. However, the relationsh...
详细信息
Motions are important features for robot vision as we live in a dynamic world. Detecting moving objects is crucial for mobile robots and computer vision systems. This paper investigates an architecture for the segment...
详细信息
Motions are important features for robot vision as we live in a dynamic world. Detecting moving objects is crucial for mobile robots and computer vision systems. This paper investigates an architecture for the segmentation of moving objects from image sequences. Objects are represented as groups of SIFT feature points. Instead of tracking the feature points over a sequence of frames, the movements of feature points between two successive frames are used. The segmentation of motions of each pair of frames is based on the expectation-maximization algorithm. The segmentation algorithm is iteratively applied over all frames of the sequence and the results are combined using Bayesian update.
Convolutional Neural Networks (CNNs) are widely used in NLP tasks. This paper presents a novel weight initialization method to improve the CNNs for text classification. Instead of randomly initializing the convolution...
详细信息
The Endoscopy Computer Vision Challenge (EndoCV) is a crowd-sourcing initiative to address eminent problems in developing reliable computer aided detection and diagnosis endoscopy systems and suggest a pathway for cli...
详细信息
We examine the long-run behavior of multi-agent online learning in games that evolve over time. Specifically, we focus on a wide class of policies based on mirror descent, and we show that the induced sequence of play...
详细信息
There’re many effective architectures of the artificial neural network(ANN). For which the training is a hard work. The cost for training an ANN increases exponentially when the ANN gets deeper or wider. We therefore...
详细信息
There’re many effective architectures of the artificial neural network(ANN). For which the training is a hard work. The cost for training an ANN increases exponentially when the ANN gets deeper or wider. We therefore propose a novel architecture, the Hybrid Learning Network(HLN), to achieve a fast learning with good stablity. The HLN can learn from both labeled data and unlabeled data at the same time in a hybrid learning manner. It uses a Self Organizing Map unified by the specially designed nonlinear function as the sparsity mask for a hidden layer to improve the training speed. We experiment our architecture on a synthetic dataset to test its regression capability against the traditional architecture, the result is promising.
Feature selection based on information theory plays an important role in classification algorithm due to its computational efficiency and independent from classification method. It is widely used in many application a...
详细信息
Feature selection based on information theory plays an important role in classification algorithm due to its computational efficiency and independent from classification method. It is widely used in many application areas like data mining, bioinformatics and machine learning. But drawbacks of these methods are the neglect of the feature interaction and overestimation of features significance due to the limitations of goal functions criterion. To address this problem, we proposed a new feature goal function RJMIM. The method employed joint mutual information and information interaction, which alleviates the shortcomings of overestimation of the feature significance as demonstrated both theoretically and experimentally. The experiments conducted to verify the performance of the proposed method, it compared with four well-known feature selection methods use three publically available datasets from UCI. The average classification accuracy and C4.5 classifier is used to assess the effectiveness of RJMIM method.
暂无评论