Multiple kernel learning

About: Multiple kernel learning is a research topic. Over the lifetime, 1630 publications have been published within this topic receiving 56082 citations.

Composite Kernel of Mutual Learning on Mid-Level Features for Hyperspectral Image Classification

Abstract: By training different models and averaging their predictions, the performance of the machine-learning algorithm can be improved. The performance optimization of multiple models is supposed to generalize further data well. This requires the knowledge transfer of generalization information between models. In this article, a multiple kernel mutual learning method based on transfer learning of combined mid-level features is proposed for hyperspectral classification. Three-layer homogenous superpixels are computed on the image formed by PCA, which is used for computing mid-level features. The three mid-level features include: 1) the sparse reconstructed feature; 2) combined mean feature; and 3) uniqueness. The sparse reconstruction feature is obtained by a joint sparse representation model under the constraint of three-scale superpixels' boundaries and regions. The combined mean features are computed with average values of spectra in multilayer superpixels, and the uniqueness is obtained by the superposed manifold ranking values of multilayer superpixels. Next, three kernels of samples in different feature spaces are computed for mutual learning by minimizing the divergence. Then, a combined kernel is constructed to optimize the sample distance measurement and applied by employing SVM training to build classifiers. Experiments are performed on real hyperspectral datasets, and the corresponding results demonstrated that the proposed method can perform significantly better than several state-of-the-art competitive algorithms based on MKL and deep learning.

An adaboost-based weighting method for localizing human brain magnetic activity

Abstract: This paper shows that pattern classification based on machine learning is a powerful tool for analyzing human brain activity data obtained by magnetoencephalography (MEG) In our previous work, a weighting method using multiple kernel learning was proposed, but this method had a high computational cost In this paper, we propose a novel and fast weighting method using an AdaBoost algorithm to find the sensor area contributing to the accurate discrimination of vowels Our AdaBoost simultaneously estimates both the classification boundary and the weight to each MEG sensor, with MEG amplitude obtained from each pair of sensors being an element of the feature vector The estimated weight indicates how the corresponding sensor is useful for classifying the MEG response patterns Our results for vowel recognition show the large-weight MEG sensors mainly in a language area of the brain and the high classification accuracy (910%) in the latency range between 50 and 150 ms

A Multiple Kernel Learning Framework to Investigate the Relationship Between Ventricular Fibrillation and First Myocardial Infarction

Abstract: Myocardial infarction results in changes in the structure and tissue deformation of the ventricles. In some cases, the development of the disease may trigger an arrhythmic event, which is a major cause of death within the first twenty four hours after the infarction. Advanced analysis methods are increasingly used in order to discover particular characteristics of the myocardial infarction development that lead to the occurrence of arrhythmias. However, such methods usually consider only a single feature or combine separate analyses from multiple features in the analytical process. In an attempt to address this, we propose to use cardiac magnetic resonance imaging to extract data on the shape of the ventricles and volume and location of the infarct zone, and to combine them within one analytical model through a multiple kernel learning framework. The proposed method was applied to a cohort of 46 myocardial infarction patients. The location, rather than the volume, of the infarct region was found to be correlated with arrhythmic events and the proposed combination of kernels yielded excellent accuracy (100%) in distinguishing between patients that did and did not present at the hospital with ventricular fibrillation.

Question Classification using Multiple Kernel Learning and Semantic Information

Abstract: Question Classification is an important stage in Question Answering, and it has been a hot topic in the field of Information Retrieval in recent years. In this paper we explore the role of semantic features and propose two separate tree kernel functions incorporating the semantic features into the Support Vector Machine model. Then Multiple Kernel Learning approach is proposed to combine the two kernels and gather their advantages together. Experimental results show that using the method proposed in this paper is very effective and the accuracy reaches 95.8% which significantly outperforms the state-of-the-art approaches.