Dictionary-based Representation Research Articles

Data-Driven Outage Restoration Time Prediction via Transfer Learning With Cluster Ensembles

This paper develops a data-driven approach to accurately predict the restoration time of outages under different scales and factors. To achieve the goal, the proposed method consists of three stages. First, given the unprecedented amount of data collected by utilities, a sparse dictionary-based ensemble spectral clustering (SDESC) method is proposed to decompose historical outage datasets, which enjoys good computational efficiency and scalability. Specifically, each outage sample is represented by a linear combination of a small number of selected dictionary samples using a density-based method. Then, the dictionary-based representation is utilized to perform the spectral analysis to group the data samples with similar features into the same subsets. In the second stage, a knowledge-transfer-added restoration time prediction model is trained for each subset by combining weather information and outage-related features. The transfer learning technology is introduced to deal with the underestimation problem caused by data imbalance in different subsets, thus improving the model performance. Furthermore, to connect unseen outages with the learned outage subsets, a t-distributed stochastic neighbor embedding-based strategy is applied. The proposed method fully builds on and is also tested on a large real-world outage dataset from a utility provider with a time span of six consecutive years. The numerical results validate that our method has high prediction accuracy while showing good stability against real-world data limitations.

Blind Source Separation in Polyphonic Music Recordings Using Deep Neural Networks Trained via Policy Gradients

We propose a method for the blind separation of sounds of musical instruments in audio signals. We describe the individual tones via a parametric model, training a dictionary to capture the relative amplitudes of the harmonics. The model parameters are predicted via a U-Net, which is a type of deep neural network. The network is trained without ground truth information, based on the difference between the model prediction and the individual time frames of the short-time Fourier transform. Since some of the model parameters do not yield a useful backpropagation gradient, we model them stochastically and employ the policy gradient instead. To provide phase information and account for inaccuracies in the dictionary-based representation, we also let the network output a direct prediction, which we then use to resynthesize the audio signals for the individual instruments. Due to the flexibility of the neural network, inharmonicity can be incorporated seamlessly and no preprocessing of the input spectra is required. Our algorithm yields high-quality separation results with particularly low interference on a variety of different audio samples, both acoustic and synthetic, provided that the sample contains enough data for the training and that the spectral characteristics of the musical instruments are sufficiently stable to be approximated by the dictionary.

Sparse representation-based classification of mysticete calls.

This paper presents an automatic classification method dedicated to mysticete calls. This method relies on sparse representations which assume that mysticete calls lie in a linear subspace described by a dictionary-based representation. The classifier accounts for noise by refusing to assign the observed signal to a given class if it is not included into the linear subspace spanned by the dictionaries of mysticete calls. Rejection of noise is achieved without feature learning. In addition, the proposed method is modular in that, call classes can be appended to or removed from the classifier without requiring retraining. The classifier is easy to design since it relies on a few parameters. Experiments on five types of mysticete calls are presented. It includes Antarctic blue whale Z-calls, two types of "Madagascar" pygmy blue whale calls, fin whale 20 Hz calls and North-Pacific blue whale D-calls. On this dataset, containing 2185 calls and 15 000 noise samples, an average recall of 96.4% is obtained and 93.3% of the noise data (persistent and transient) are correctly rejected by the classifier.

S184. Sparse representation and classification of neural spikes using supervised dictionary learning

Neural spike sorting is curial for neuroscience applications. For example, spike sorting allows designing of efficient human-computer interfaces. Therefore, many investigations, aiming for accurate computer-aided algorithms for spike detection and classification, have emerged. However, many of the available algorithms require human intervention during the classification step. This study proposes decomposition and classification of spike waveforms using dictionary learning. In sparse decomposition, each spike is represented as a unique sparse combination of the dictionary basis vectors. This can lead to interesting sorting performance. Spiking activity was detected using amplitude thresholding and a 2.5-ms window around the peak value. Under the assumption that each neuron has a consistent waveform, the dictionary atoms were learned using training samples. This learned dictionary was used to get a sparse code representation for the test samples. These codes can be used to reconstruct the original spikes and therefore it provides data compression capability. Classification of the neural spikes was done using the sparse coefficients alongside with the reconstruction error. To ensure high discrimination, the dictionary was learned such that the resultant atoms have correspondence to class labels. In other words, the dictionary is structured of sub-dictionaries each corresponds to a different spiking pattern. Subsequently, an optimization problem was solved to make sure that (i) the sub-dictionaries accurately represent the training data, and (ii) the coding coefficients are sparse and minimize the reconstruction error. A testing sample was classified by sparsely coding it over each sub-dictionary. The reconstruction error for each sub-dictionary is computed. An error function is defined for each class as the summation of the reconstruction error and the difference between the sparse code and that class average. The testing sample was assigned to the class that gave the minimum error. The algorithm was tested on four simulated data sets, each contains spiking activities from three different neurons. Approximately 10% of each class data was used to construct the dictionary. The reconstruction mean-squared error for all the training samples was 3.49. This confirms the reconstruction power and accuracy of the learned dictionary. We measured the accuracy of our algorithm at different noise levels. The classification accuracy of the proposed method is 97% ± 5%. This accuracy outperforms the existing state-of-the-art algorithms. A dictionary learning based spike sorting algorithm has been developed. The results suggest that dictionary-based representation enables us to use the sparse representation of signals without a severe loss of data; besides that, this sparse representation can be also used for classification. This could also enable estimation of missing data points or clipped signals. The reconstruction error can be used to efficiently sort spikes.

Image-based surrogate biomarkers for molecular subtypes of colorectal cancer.

Whole genome expression profiling of large cohorts of different types of cancer led to the identification of distinct molecular subcategories (subtypes) that may partially explain the observed inter-tumoral heterogeneity. This is also the case of colorectal cancer (CRC) where several such categorizations have been proposed. Despite recent developments, the problem of subtype definition and recognition remains open, one of the causes being the intrinsic heterogeneity of each tumor, which is difficult to estimate from gene expression profiles. However, one of the observations of these studies indicates that there may be links between the dominant tumor morphology characteristics and the molecular subtypes. Benefiting from a large collection of CRC samples, comprising both gene expression and histopathology images, we investigated the possibility of building image-based classifiers able to predict the molecular subtypes. We employed deep convolutional neural networks for extracting local descriptors which were then used for constructing a dictionary-based representation of each tumor sample. A set of support vector machine classifiers were trained to solve different binary decision problems, their combined outputs being used to predict one of the five molecular subtypes. A hierarchical decomposition of the multi-class problem was obtained with an overall accuracy of 0.84 (95%CI=0.79-0.88). The predictions from the image-based classifier showed significant prognostic value similar to their molecular counterparts. popovici@iba.muni.cz. Source code used for the image analysis is freely available from https://github.com/higex/qpath . Supplementary data are available at Bioinformatics online.

Whence the Unicorn? A Frame-Semantic and Corpus Linguistic Interpretation of the Notion of Unicornhood in the Characterology of Iris Murdoch’s Novel: Th

The notion of unicornhood in the characterology of Iris Murdoch’s Gothic fiction titled The Unicorn has been interpreted variously by literary critics since the publication of the novel.The present paper provides a linguistically and cultural philosophically grounded, frame-semantic analysis, interpretation as well as evidence of the concept of unicornhood, wherein behavioral and appearance-related socio-cultural stereotypes mentally stored in encyclopedically-based knowledge structures about the factually non-existent, imaginary (or, perhaps, existent) being are identified and discussed. The semantically and lexical-pragmatically based factors revealed in the analysis are supported by an empirical investigation of corpus-based evidence gained from the British National Corpus of 100 million words, by studying the dictionary-based representation of the lexical item unicorn, and also by testing the tacitly represented background knowledge of 40 native speakers of English, with the aim to provide a joint control over the corpus-based data. As a result, it is pointed out that the standard characterology offered by literary scholars requires further re-evaluation and reinterpretatrion, expectably revealing the multiple realization of unicornhood symbolism represented in the novel.

Dictionary-Driven Ischemia Detection From Cardiac Phase-Resolved Myocardial BOLD MRI at Rest.

Cardiac Phase-resolved Blood-Oxygen-Level Dependent (CP-BOLD) MRI provides a unique opportunity to image an ongoing ischemia at rest. However, it requires post-processing to evaluate the extent of ischemia. To address this, here we propose an unsupervised ischemia detection (UID) method which relies on the inherent spatio-temporal correlation between oxygenation and wall motion to formalize a joint learning and detection problem based on dictionary decomposition. Considering input data of a single subject, it treats ischemia as an anomaly and iteratively learns dictionaries to represent only normal observations (corresponding to myocardial territories remote to ischemia). Anomaly detection is based on a modified version of One-class Support Vector Machines (OCSVM) to regulate directly the margins by incorporating the dictionary-based representation errors. A measure of ischemic extent (IE) is estimated, reflecting the relative portion of the myocardium affected by ischemia. For visualization purposes an ischemia likelihood map is created by estimating posterior probabilities from the OCSVM outputs, thus obtaining how likely the classification is correct. UID is evaluated on synthetic data and in a 2D CP-BOLD data set from a canine experimental model emulating acute coronary syndromes. Comparing early ischemic territories identified with UID against infarct territories (after several hours of ischemia), we find that IE, as measured by UID, is highly correlated (Pearson's r=0.84) with respect to infarct size. When advances in automated registration and segmentation of CP-BOLD images and full coverage 3D acquisitions become available, we hope that this method can enable pixel-level assessment of ischemia with this truly non-invasive imaging technique.

Dictionary-based image reconstruction for superresolution in integrated circuit imaging.

Resolution improvement through signal processing techniques for integrated circuit imaging is becoming more crucial as the rapid decrease in integrated circuit dimensions continues. Although there is a significant effort to push the limits of optical resolution for backside fault analysis through the use of solid immersion lenses, higher order laser beams, and beam apodization, signal processing techniques are required for additional improvement. In this work, we propose a sparse image reconstruction framework which couples overcomplete dictionary-based representation with a physics-based forward model to improve resolution and localization accuracy in high numerical aperture confocal microscopy systems for backside optical integrated circuit analysis. The effectiveness of the framework is demonstrated on experimental data.