L1-Norm Distance Discriminant Analysis with Multiple Adaptive Graphs and Sample Reconstruction

Linear discriminant analysis (LDA) is sensitive to noise and its performance may decline greatly. Recursive discriminative subspace learning method with an L1-norm distance constraint (RDSL) formulates LDA with the maximum margin criterion and becomes robust to noise by applying L1-norm and slack variables. However, the method only considers inter-class separation and intra-class compactness and ignores the intra-class manifold structure and the global structure of data. In this paper, we present L1-norm distance discriminant analysis with multiple adaptive graphs and sample reconstruction (L1-DDA) to deal with the problem. We use multiple adaptive graphs to preserve intra-class manifold structure and simultaneously apply the sample reconstruction technique to preserve the global structure of data. Moreover, we use an alternating iterative technique to obtain projection vectors. Experimental results on three real databases demonstrate that our method obtains better classification performance than RDSL.

Cutting L1-Norm Distance Discriminant Analysis with Sample Reconstruction

Dimensionality reduction plays an important role in the fields of pattern recognition and computer vision. Recursive discriminative subspace learning with an L1-norm distance constraint (RDSL) is proposed to robustly extract features from contaminated data and L1-norm and slack variables are utilized for accomplishing the goal. However, its performance may decline when too many outliers are available. Moreover, the method ignores the global structure of the data. In this paper, we propose cutting L1-norm distance discriminant analysis with sample reconstruction (C-L1-DDA) to solve the two problems. We apply cutting L1-norm to measure within-class and between-class distances and thus outliers may be strongly suppressed. Moreover, we use cutting squared L2-norm to measure reconstruction errors. In this way, outliers may be constrained and the global structure of data may be approximately preserved. Finally, we give an alternating iterative algorithm to extract feature vectors. Experimental results on two publicly available real databases verify the feasibility and effectiveness of the proposed method.

Transfer Learning and Deep Domain Adaptation

Transfer learning is an emerging technique in machine learning, by which we can solve a new task with the knowledge obtained from an old task in order to address the lack of labeled data. In particular deep domain adaptation (a branch of transfer learning) gets the most attention in recently published articles. The intuition behind this is that deep neural networks usually have a large capacity to learn representation from one dataset and part of the information can be further used for a new task. In this research, we firstly present the complete scenarios of transfer learning according to the domains and tasks. Secondly, we conduct a comprehensive survey related to deep domain adaptation and categorize the recent advances into three types based on implementing approaches: fine-tuning networks, adversarial domain adaptation, and sample-reconstruction approaches. Thirdly, we discuss the details of these methods and introduce some typical real-world applications. Finally, we conclude our work and explore some potential issues to be further addressed.

Effect of radiation damage and illumination variability on signal-to-noise ratio in X-ray free-electron laser single-particle imaging

The deterioration of both the signal-to-noise ratio and the spatial resolution in the electron-density distribution reconstructed from diffraction intensities collected at different orientations of a sample is analysed theoretically with respect to the radiation damage to the sample and the variations in the X-ray intensities illuminating different copies of the sample. The simple analytical expressions and numerical estimates obtained for models of radiation damage and incident X-ray pulses may be helpful in planning X-ray free-electron laser (XFEL) imaging experiments and in analysis of experimental data. This approach to the analysis of partially coherent X-ray imaging configurations can potentially be used for analysis of other forms of imaging where the temporal behaviour of the sample and the incident intensity during exposure may affect the inverse problem of sample reconstruction.