Search results for: sparse dictionary learning
Commenced in January 2007
Frequency: Monthly
Edition: International
Paper Count: 6948

Search results for: sparse dictionary learning

6948 KSVD-SVM Approach for Spontaneous Facial Expression Recognition

Authors: Dawood Al Chanti, Alice Caplier

Abstract:

Sparse representations of signals have received a great deal of attention in recent years. In this paper, the interest of using sparse representation as a mean for performing sparse discriminative analysis between spontaneous facial expressions is demonstrated. An automatic facial expressions recognition system is presented. It uses a KSVD-SVM approach which is made of three main stages: A pre-processing and feature extraction stage, which solves the problem of shared subspace distribution based on the random projection theory, to obtain low dimensional discriminative and reconstructive features; A dictionary learning and sparse coding stage, which uses the KSVD model to learn discriminative under or over dictionaries for sparse coding; Finally a classification stage, which uses a SVM classifier for facial expressions recognition. Our main concern is to be able to recognize non-basic affective states and non-acted expressions. Extensive experiments on the JAFFE static acted facial expressions database but also on the DynEmo dynamic spontaneous facial expressions database exhibit very good recognition rates.

Keywords: dictionary learning, random projection, pose and spontaneous facial expression, sparse representation

Procedia PDF Downloads 264
6947 Sparse Coding Based Classification of Electrocardiography Signals Using Data-Driven Complete Dictionary Learning

Authors: Fuad Noman, Sh-Hussain Salleh, Chee-Ming Ting, Hadri Hussain, Syed Rasul

Abstract:

In this paper, a data-driven dictionary approach is proposed for the automatic detection and classification of cardiovascular abnormalities. Electrocardiography (ECG) signal is represented by the trained complete dictionaries that contain prototypes or atoms to avoid the limitations of pre-defined dictionaries. The data-driven trained dictionaries simply take the ECG signal as input rather than extracting features to study the set of parameters that yield the most descriptive dictionary. The approach inherently learns the complicated morphological changes in ECG waveform, which is then used to improve the classification. The classification performance was evaluated with ECG data under two different preprocessing environments. In the first category, QT-database is baseline drift corrected with notch filter and it filters the 60 Hz power line noise. In the second category, the data are further filtered using fast moving average smoother. The experimental results on QT database confirm that our proposed algorithm shows a classification accuracy of 92%.

Keywords: electrocardiogram, dictionary learning, sparse coding, classification

Procedia PDF Downloads 342
6946 Sparse Representation Based Spatiotemporal Fusion Employing Additional Image Pairs to Improve Dictionary Training

Authors: Dacheng Li, Bo Huang, Qinjin Han, Ming Li

Abstract:

Remotely sensed imagery with the high spatial and temporal characteristics, which it is hard to acquire under the current land observation satellites, has been considered as a key factor for monitoring environmental changes over both global and local scales. On a basis of the limited high spatial-resolution observations, challenged studies called spatiotemporal fusion have been developed for generating high spatiotemporal images through employing other auxiliary low spatial-resolution data while with high-frequency observations. However, a majority of spatiotemporal fusion approaches yield to satisfactory assumption, empirical but unstable parameters, low accuracy or inefficient performance. Although the spatiotemporal fusion methodology via sparse representation theory has advantage in capturing reflectance changes, stability and execution efficiency (even more efficient when overcomplete dictionaries have been pre-trained), the retrieval of high-accuracy dictionary and its response to fusion results are still pending issues. In this paper, we employ additional image pairs (here each image-pair includes a Landsat Operational Land Imager and a Moderate Resolution Imaging Spectroradiometer acquisitions covering the partial area of Baotou, China) only into the coupled dictionary training process based on K-SVD (K-means Singular Value Decomposition) algorithm, and attempt to improve the fusion results of two existing sparse representation based fusion models (respectively utilizing one and two available image-pair). The results show that more eligible image pairs are probably related to a more accurate overcomplete dictionary, which generally indicates a better image representation, and is then contribute to an effective fusion performance in case that the added image-pair has similar seasonal aspects and image spatial structure features to the original image-pair. It is, therefore, reasonable to construct multi-dictionary training pattern for generating a series of high spatial resolution images based on limited acquisitions.

Keywords: spatiotemporal fusion, sparse representation, K-SVD algorithm, dictionary learning

Procedia PDF Downloads 218
6945 Atomic Decomposition Audio Data Compression and Denoising Using Sparse Dictionary Feature Learning

Authors: T. Bryan , V. Kepuska, I. Kostnaic

Abstract:

A method of data compression and denoising is introduced that is based on atomic decomposition of audio data using “basis vectors” that are learned from the audio data itself. The basis vectors are shown to have higher data compression and better signal-to-noise enhancement than the Gabor and gammatone “seed atoms” that were used to generate them. The basis vectors are the input weights of a Sparse AutoEncoder (SAE) that is trained using “envelope samples” of windowed segments of the audio data. The envelope samples are extracted from the audio data by performing atomic decomposition with Gabor or gammatone seed atoms. This process identifies segments of audio data that are locally coherent with the seed atoms. Envelope samples are extracted by identifying locally coherent audio data segments with Gabor or gammatone seed atoms, found by matching pursuit. The envelope samples are formed by taking the kronecker products of the atomic envelopes with the locally coherent data segments. Oracle signal-to-noise ratio (SNR) verses data compression curves are generated for the seed atoms as well as the basis vectors learned from Gabor and gammatone seed atoms. SNR data compression curves are generated for speech signals as well as early American music recordings. The basis vectors are shown to have higher denoising capability for data compression rates ranging from 90% to 99.84% for speech as well as music. Envelope samples are displayed as images by folding the time series into column vectors. This display method is used to compare of the output of the SAE with the envelope samples that produced them. The basis vectors are also displayed as images. Sparsity is shown to play an important role in producing the highest denoising basis vectors.

Keywords: sparse dictionary learning, autoencoder, sparse autoencoder, basis vectors, atomic decomposition, envelope sampling, envelope samples, Gabor, gammatone, matching pursuit

Procedia PDF Downloads 217
6944 Speaker Identification by Atomic Decomposition of Learned Features Using Computational Auditory Scene Analysis Principals in Noisy Environments

Authors: Thomas Bryan, Veton Kepuska, Ivica Kostanic

Abstract:

Speaker recognition is performed in high Additive White Gaussian Noise (AWGN) environments using principals of Computational Auditory Scene Analysis (CASA). CASA methods often classify sounds from images in the time-frequency (T-F) plane using spectrograms or cochleargrams as the image. In this paper atomic decomposition implemented by matching pursuit performs a transform from time series speech signals to the T-F plane. The atomic decomposition creates a sparsely populated T-F vector in “weight space” where each populated T-F position contains an amplitude weight. The weight space vector along with the atomic dictionary represents a denoised, compressed version of the original signal. The arraignment or of the atomic indices in the T-F vector are used for classification. Unsupervised feature learning implemented by a sparse autoencoder learns a single dictionary of basis features from a collection of envelope samples from all speakers. The approach is demonstrated using pairs of speakers from the TIMIT data set. Pairs of speakers are selected randomly from a single district. Each speak has 10 sentences. Two are used for training and 8 for testing. Atomic index probabilities are created for each training sentence and also for each test sentence. Classification is performed by finding the lowest Euclidean distance between then probabilities from the training sentences and the test sentences. Training is done at a 30dB Signal-to-Noise Ratio (SNR). Testing is performed at SNR’s of 0 dB, 5 dB, 10 dB and 30dB. The algorithm has a baseline classification accuracy of ~93% averaged over 10 pairs of speakers from the TIMIT data set. The baseline accuracy is attributable to short sequences of training and test data as well as the overall simplicity of the classification algorithm. The accuracy is not affected by AWGN and produces ~93% accuracy at 0dB SNR.

Keywords: time-frequency plane, atomic decomposition, envelope sampling, Gabor atoms, matching pursuit, sparse dictionary learning, sparse autoencoder

Procedia PDF Downloads 250
6943 The Efficiency of the Use of Medical Bilingual Dictionary in English Language Teaching in Vocational College

Authors: Zorana Jurinjak, Christos Alexopoulos

Abstract:

The aim of this paper is to examine the effectiveness of using a medical bilingual dictionary in teaching English in a vocational college. More precisely, to what extent the use of bilingual medical dictionary in relation to the use of Standard English bilingual dictionaries influences the results on tests, and thus the acquisition of better competence of students mastering the subject terminology. Secondary interest in this paper would be to raise awareness among students and teachers about the advantages of dictionary use. The experiment was conducted at College of Applied Health Sciences in Ćuprija on a sample of 90 students. The respondents translated three medical texts with 42 target terms. Statistical analyses of the data obtained show that the differences in average time and correct answers favor the students who used medical dictionary.

Keywords: bilingual medical dictionary, standard english bilingual dictionary, medical terminology, EOS, ESP

Procedia PDF Downloads 74
6942 The Grammatical Dictionary Compiler: A System for Kartvelian Languages

Authors: Liana Lortkipanidze, Nino Amirezashvili, Nino Javashvili

Abstract:

The purpose of the grammatical dictionary is to provide information on the morphological and syntactic characteristics of the basic word in the dictionary entry. The electronic grammatical dictionaries are used as a tool of automated morphological analysis for texts processing. The Georgian Grammatical Dictionary should contain grammatical information for each word: part of speech, type of declension/conjugation, grammatical forms of the word (paradigm), alternative variants of basic word/lemma. In this paper, we present the system for compiling the Georgian Grammatical Dictionary automatically. We propose dictionary-based methods for extending grammatical lexicons. The input lexicon contains only a few number of words with identical grammatical features. The extension is based on similarity measures between features of words; more precisely, we add words to the extended lexicons, which are similar to those, which are already in the grammatical dictionary. Our dictionaries are corpora-based, and for the compiling, we introduce the method for lemmatization of unknown words, i.e., words of which neither full form nor lemma is in the grammatical dictionary.

Keywords: acquisition of lexicon, Georgian grammatical dictionary, lemmatization rules, morphological processor

Procedia PDF Downloads 107
6941 An Image Segmentation Algorithm for Gradient Target Based on Mean-Shift and Dictionary Learning

Authors: Yanwen Li, Shuguo Xie

Abstract:

In electromagnetic imaging, because of the diffraction limited system, the pixel values could change slowly near the edge of the image targets and they also change with the location in the same target. Using traditional digital image segmentation methods to segment electromagnetic gradient images could result in lots of errors because of this change in pixel values. To address this issue, this paper proposes a novel image segmentation and extraction algorithm based on Mean-Shift and dictionary learning. Firstly, the preliminary segmentation results from adaptive bandwidth Mean-Shift algorithm are expanded, merged and extracted. Then the overlap rate of the extracted image block is detected before determining a segmentation region with a single complete target. Last, the gradient edge of the extracted targets is recovered and reconstructed by using a dictionary-learning algorithm, while the final segmentation results are obtained which are very close to the gradient target in the original image. Both the experimental results and the simulated results show that the segmentation results are very accurate. The Dice coefficients are improved by 70% to 80% compared with the Mean-Shift only method.

Keywords: gradient image, segmentation and extract, mean-shift algorithm, dictionary iearning

Procedia PDF Downloads 228
6940 A Sparse Representation Speech Denoising Method Based on Adapted Stopping Residue Error

Authors: Qianhua He, Weili Zhou, Aiwu Chen

Abstract:

A sparse representation speech denoising method based on adapted stopping residue error was presented in this paper. Firstly, the cross-correlation between the clean speech spectrum and the noise spectrum was analyzed, and an estimation method was proposed. In the denoising method, an over-complete dictionary of the clean speech power spectrum was learned with the K-singular value decomposition (K-SVD) algorithm. In the sparse representation stage, the stopping residue error was adaptively achieved according to the estimated cross-correlation and the adjusted noise spectrum, and the orthogonal matching pursuit (OMP) approach was applied to reconstruct the clean speech spectrum from the noisy speech. Finally, the clean speech was re-synthesised via the inverse Fourier transform with the reconstructed speech spectrum and the noisy speech phase. The experiment results show that the proposed method outperforms the conventional methods in terms of subjective and objective measure.

Keywords: speech denoising, sparse representation, k-singular value decomposition, orthogonal matching pursuit

Procedia PDF Downloads 464
6939 Performance Analysis and Optimization for Diagonal Sparse Matrix-Vector Multiplication on Machine Learning Unit

Authors: Qiuyu Dai, Haochong Zhang, Xiangrong Liu

Abstract:

Diagonal sparse matrix-vector multiplication is a well-studied topic in the fields of scientific computing and big data processing. However, when diagonal sparse matrices are stored in DIA format, there can be a significant number of padded zero elements and scattered points, which can lead to a degradation in the performance of the current DIA kernel. This can also lead to excessive consumption of computational and memory resources. In order to address these issues, the authors propose the DIA-Adaptive scheme and its kernel, which leverages the parallel instruction sets on MLU. The researchers analyze the effect of allocating a varying number of threads, clusters, and hardware architectures on the performance of SpMV using different formats. The experimental results indicate that the proposed DIA-Adaptive scheme performs well and offers excellent parallelism.

Keywords: adaptive method, DIA, diagonal sparse matrices, MLU, sparse matrix-vector multiplication

Procedia PDF Downloads 67
6938 Music Note Detection and Dictionary Generation from Music Sheet Using Image Processing Techniques

Authors: Muhammad Ammar, Talha Ali, Abdul Basit, Bakhtawar Rajput, Zobia Sohail

Abstract:

Music note detection is an area of study for the past few years and has its own influence in music file generation from sheet music. We proposed a method to detect music notes on sheet music using basic thresholding and blob detection. Subsequently, we created a notes dictionary using a semi-supervised learning approach. After notes detection, for each test image, the new symbols are added to the dictionary. This makes the notes detection semi-automatic. The experiments are done on images from a dataset and also on the captured images. The developed approach showed almost 100% accuracy on the dataset images, whereas varying results have been seen on captured images.

Keywords: music note, sheet music, optical music recognition, blob detection, thresholding, dictionary generation

Procedia PDF Downloads 131
6937 Non-Local Simultaneous Sparse Unmixing for Hyperspectral Data

Authors: Fanqiang Kong, Chending Bian

Abstract:

Sparse unmixing is a promising approach in a semisupervised fashion by assuming that the observed pixels of a hyperspectral image can be expressed in the form of linear combination of only a few pure spectral signatures (end members) in an available spectral library. However, the sparse unmixing problem still remains a great challenge at finding the optimal subset of endmembers for the observed data from a large standard spectral library, without considering the spatial information. Under such circumstances, a sparse unmixing algorithm termed as non-local simultaneous sparse unmixing (NLSSU) is presented. In NLSSU, the non-local simultaneous sparse representation method for endmember selection of sparse unmixing, is used to finding the optimal subset of endmembers for the similar image patch set in the hyperspectral image. And then, the non-local means method, as a regularizer for abundance estimation of sparse unmixing, is used to exploit the abundance image non-local self-similarity. Experimental results on both simulated and real data demonstrate that NLSSU outperforms the other algorithms, with a better spectral unmixing accuracy.

Keywords: hyperspectral unmixing, simultaneous sparse representation, sparse regression, non-local means

Procedia PDF Downloads 199
6936 Curriculum-Based Multi-Agent Reinforcement Learning for Robotic Navigation

Authors: Hyeongbok Kim, Lingling Zhao, Xiaohong Su

Abstract:

Deep reinforcement learning has been applied to address various problems in robotics, such as autonomous driving and unmanned aerial vehicle. However, because of the sparse reward penalty for a collision with obstacles during the navigation mission, the agent fails to learn the optimal policy or requires a long time for convergence. Therefore, using obstacles and enemy agents, in this paper, we present a curriculum-based boost learning method to effectively train compound skills during multi-agent reinforcement learning. First, to enable the agents to solve challenging tasks, we gradually increased learning difficulties by adjusting reward shaping instead of constructing different learning environments. Then, in a benchmark environment with static obstacles and moving enemy agents, the experimental results showed that the proposed curriculum learning strategy enhanced cooperative navigation and compound collision avoidance skills in uncertain environments while improving learning efficiency.

Keywords: curriculum learning, hard exploration, multi-agent reinforcement learning, robotic navigation, sparse reward

Procedia PDF Downloads 55
6935 An Improved Method to Compute Sparse Graphs for Traveling Salesman Problem

Authors: Y. Wang

Abstract:

The Traveling salesman problem (TSP) is NP-hard in combinatorial optimization. The research shows the algorithms for TSP on the sparse graphs have the shorter computation time than those for TSP according to the complete graphs. We present an improved iterative algorithm to compute the sparse graphs for TSP by frequency graphs computed with frequency quadrilaterals. The iterative algorithm is enhanced by adjusting two parameters of the algorithm. The computation time of the algorithm is O(CNmaxn2) where C is the iterations, Nmax is the maximum number of frequency quadrilaterals containing each edge and n is the scale of TSP. The experimental results showed the computed sparse graphs generally have less than 5n edges for most of these Euclidean instances. Moreover, the maximum degree and minimum degree of the vertices in the sparse graphs do not have much difference. Thus, the computation time of the methods to resolve the TSP on these sparse graphs will be greatly reduced.

Keywords: frequency quadrilateral, iterative algorithm, sparse graph, traveling salesman problem

Procedia PDF Downloads 180
6934 A Generalized Sparse Bayesian Learning Algorithm for Near-Field Synthetic Aperture Radar Imaging: By Exploiting Impropriety and Noncircularity

Authors: Pan Long, Bi Dongjie, Li Xifeng, Xie Yongle

Abstract:

The near-field synthetic aperture radar (SAR) imaging is an advanced nondestructive testing and evaluation (NDT&E) technique. This paper investigates the complex-valued signal processing related to the near-field SAR imaging system, where the measurement data turns out to be noncircular and improper, meaning that the complex-valued data is correlated to its complex conjugate. Furthermore, we discover that the degree of impropriety of the measurement data and that of the target image can be highly correlated in near-field SAR imaging. Based on these observations, A modified generalized sparse Bayesian learning algorithm is proposed, taking impropriety and noncircularity into account. Numerical results show that the proposed algorithm provides performance gain, with the help of noncircular assumption on the signals.

Keywords: complex-valued signal processing, synthetic aperture radar, 2-D radar imaging, compressive sensing, sparse Bayesian learning

Procedia PDF Downloads 87
6933 Analysis of the Significance of Multimedia Channels Using Sparse PCA and Regularized SVD

Authors: Kourosh Modarresi

Abstract:

The abundance of media channels and devices has given users a variety of options to extract, discover, and explore information in the digital world. Since, often, there is a long and complicated path that a typical user may venture before taking any (significant) action (such as purchasing goods and services), it is critical to know how each node (media channel) in the path of user has contributed to the final action. In this work, the significance of each media channel is computed using statistical analysis and machine learning techniques. More specifically, “Regularized Singular Value Decomposition”, and “Sparse Principal Component” has been used to compute the significance of each channel toward the final action. The results of this work are a considerable improvement compared to the present approaches.

Keywords: multimedia attribution, sparse principal component, regularization, singular value decomposition, feature significance, machine learning, linear systems, variable shrinkage

Procedia PDF Downloads 272
6932 A Comparative Analysis of Vocabulary Learning Strategies among EFL Freshmen and Senior Medical Sciences Students across Different Fields of Study

Authors: M. Hadavi, Z. Hashemi

Abstract:

Learning strategies play an important role in the development of language skills. Vocabulary learning strategies as the backbone of these strategies have become a major part of English language teaching. This study is a comparative analysis of Vocabulary Learning Strategies (VLS) use and preference among freshmen and senior EFL medical sciences students with different fields of study. 449 students (236 freshman and 213 seniors) participated in the study. 64.6% were female and 35.4% were male. The instrument utilized in this research was a questionnaire consisting of 41 items related to the students’ approach to vocabulary learning. The items were classified under eight sections as dictionary strategies, guessing strategies, study preferences, memory strategies, autonomy, note- taking strategies, selective attention, and social strategies. The participants were asked to answer each item with a 5-point Likert-style frequency scale as follows:1) I never or almost never do this, 2) I don’t usually do this, 3) I sometimes do this, 4) I usually do this, and 5)I always or almost always do this. The results indicated that freshmen students and particularly surgical technology students used more strategies compared to the seniors. Overall guessing and dictionary strategies were the most frequently used strategies among all the learners (p=0/000). The mean and standard deviation of using VLS in the students who had no previous history of participating in the private English language classes was less than the students who had attended these type of classes (p=0/000). Female students tended to use social and study preference strategies whereas male students used mostly guessing and dictionary strategies. It can be concluded that the senior students under instruction from the university have learned to rely on themselves and choose the autonomous strategies more, while freshmen students use more strategies that are related to the study preferences.

Keywords: vocabulary leaning strategies, medical sciences, students, linguistics

Procedia PDF Downloads 407
6931 Scalable Learning of Tree-Based Models on Sparsely Representable Data

Authors: Fares Hedayatit, Arnauld Joly, Panagiotis Papadimitriou

Abstract:

Many machine learning tasks such as text annotation usually require training over very big datasets, e.g., millions of web documents, that can be represented in a sparse input space. State-of the-art tree-based ensemble algorithms cannot scale to such datasets, since they include operations whose running time is a function of the input space size rather than a function of the non-zero input elements. In this paper, we propose an efficient splitting algorithm to leverage input sparsity within decision tree methods. Our algorithm improves training time over sparse datasets by more than two orders of magnitude and it has been incorporated in the current version of scikit-learn.org, the most popular open source Python machine learning library.

Keywords: big data, sparsely representable data, tree-based models, scalable learning

Procedia PDF Downloads 224
6930 Sparse Principal Component Analysis: A Least Squares Approximation Approach

Authors: Giovanni Merola

Abstract:

Sparse Principal Components Analysis aims to find principal components with few non-zero loadings. We derive such sparse solutions by adding a genuine sparsity requirement to the original Principal Components Analysis (PCA) objective function. This approach differs from others because it preserves PCA's original optimality: uncorrelatedness of the components and least squares approximation of the data. To identify the best subset of non-zero loadings we propose a branch-and-bound search and an iterative elimination algorithm. This last algorithm finds sparse solutions with large loadings and can be run without specifying the cardinality of the loadings and the number of components to compute in advance. We give thorough comparisons with the existing sparse PCA methods and several examples on real datasets.

Keywords: SPCA, uncorrelated components, branch-and-bound, backward elimination

Procedia PDF Downloads 335
6929 The Analysis of Indian Culture through the Lexicographical Discourse of Hindi-French Dictionary

Authors: Tanzil Ansari

Abstract:

A dictionary is often considered as a list of words, arranged in alphabetical orders, providing information on a language or languages and it informs us about the spelling, the pronunciation, the origin, the gender and the grammatical functions of new and unknown words. In other words, it is first and foremost a linguistic tool. But, the research across the world in the field of linguistic and lexicography proved that a dictionary is not only a linguistic tool but also a cultural product through which a lexicographer transmits the culture of a country or a linguistic community from his or her ideology. It means, a dictionary does not present only language and its metalinguistic functions but also its culture. Every language consists of some words and expressions which depict the culture of its language. In this way, it is impossible to disassociate language from its culture. There is always an ideology that plays an important role in the depiction of any culture. Using the orientalism theory of Edward Said to represent the east, the objective of the present research is to study the representation of Indian culture through the lexicographical discourse of Hindi-French Dictionary of Federica Boschetti, a French lexicographer. The results show that the Indian culture is stereotypical and monolithic. It also shows India as male oriented country where women are exploited by male-dominated society. The study is focused on Hindi-French dictionary, but its line of argument can be compared to dictionaries produced in other languages.

Keywords: culture, dictionary, lexicographical discourse, stereotype image

Procedia PDF Downloads 262
6928 An Online Corpus-Based Bilingual Collocations Dictionary for Second/Foreign Language Learners

Authors: Adriane Orenha-Ottaiano

Abstract:

Collocations are conventionalized, recurrent and arbitrary lexical combinations. Due to the fact that they are highly specific for a particular language and may be contextually restricted, collocations pose a problem to EFL/ESL learners with regard to production or encoding. Taking that into account, the compilation of monolingual and bilingual collocations dictionaries for the referred audience is highly crucial and significant. Thus, the aim of this paper is to discuss the importance of the compilation of an Online Corpus-based Bilingual Collocations Dictionary, in the English-Portuguese and Portuguese-English directions. On a first phase, with the use of WordSmith Tools, the collocations were extracted from a Translation Learner Corpus (TLC), a parallel corpus made up of university students’ translations in the Portuguese-English direction, with approximately 100,000 words. In a second stage, based on the keywords analyzed from the TLC, more collocational patterns were extracted using the Sketch Engine. In order to include more collocations as well as to ensure dictionary users will have access to more frequent and recurrent collocations, we also use the frequency list from The Corpus of Contemporary American English, with the purpose of extracting more patterns. The dictionary focuses on all types of collocations (verbal, noun, adjectival and adverbial collocations), in order to help the referred audience use them more accurately and productively – so far the dictionary has more than 330 entries, and more than 3,500 collocations extracted. The idea of having the proposed dictionary in online format may allow to incorporate more qualitatively and quantitatively collocational information. Besides, more examples may be included, different from conventional printed collocations dictionaries. Being the first bilingual collocations dictionary in the aforementioned directions, it is hoped to achieve the challenge of meeting learners’ collocational needs as the collocations have been selected according to learners’ difficulties regarding the use of collocations.

Keywords: Corpus-Based Collocations Dictionary, Collocations , Bilingual Collocations Dictionary, Collocational Patterns

Procedia PDF Downloads 274
6927 Channel Estimation Using Deep Learning for Reconfigurable Intelligent Surfaces-Assisted Millimeter Wave Systems

Authors: Ting Gao, Mingyue He

Abstract:

Reconfigurable intelligent surfaces (RISs) are expected to be an important part of next-generation wireless communication networks due to their potential to reduce the hardware cost and energy consumption of millimeter Wave (mmWave) massive multiple-input multiple-output (MIMO) technology. However, owing to the lack of signal processing abilities of the RIS, the perfect channel state information (CSI) in RIS-assisted communication systems is difficult to acquire. In this paper, the uplink channel estimation for mmWave systems with a hybrid active/passive RIS architecture is studied. Specifically, a deep learning-based estimation scheme is proposed to estimate the channel between the RIS and the user. In particular, the sparse structure of the mmWave channel is exploited to formulate the channel estimation as a sparse reconstruction problem. To this end, the proposed approach is derived to obtain the distribution of non-zero entries in a sparse channel. After that, the channel is reconstructed by utilizing the least-squares (LS) algorithm and compressed sensing (CS) theory. The simulation results demonstrate that the proposed channel estimation scheme is superior to existing solutions even in low signal-to-noise ratio (SNR) environments.

Keywords: channel estimation, reconfigurable intelligent surface, wireless communication, deep learning

Procedia PDF Downloads 102
6926 Modern Machine Learning Conniptions for Automatic Speech Recognition

Authors: S. Jagadeesh Kumar

Abstract:

This expose presents a luculent of recent machine learning practices as employed in the modern and as pertinent to prospective automatic speech recognition schemes. The aspiration is to promote additional traverse ablution among the machine learning and automatic speech recognition factions that have transpired in the precedent. The manuscript is structured according to the chief machine learning archetypes that are furthermore trendy by now or have latency for building momentous hand-outs to automatic speech recognition expertise. The standards offered and convoluted in this article embraces adaptive and multi-task learning, active learning, Bayesian learning, discriminative learning, generative learning, supervised and unsupervised learning. These learning archetypes are aggravated and conferred in the perspective of automatic speech recognition tools and functions. This manuscript bequeaths and surveys topical advances of deep learning and learning with sparse depictions; further limelight is on their incessant significance in the evolution of automatic speech recognition.

Keywords: automatic speech recognition, deep learning methods, machine learning archetypes, Bayesian learning, supervised and unsupervised learning

Procedia PDF Downloads 404
6925 Sparsity Order Selection and Denoising in Compressed Sensing Framework

Authors: Mahdi Shamsi, Tohid Yousefi Rezaii, Siavash Eftekharifar

Abstract:

Compressed sensing (CS) is a new powerful mathematical theory concentrating on sparse signals which is widely used in signal processing. The main idea is to sense sparse signals by far fewer measurements than the Nyquist sampling rate, but the reconstruction process becomes nonlinear and more complicated. Common dilemma in sparse signal recovery in CS is the lack of knowledge about sparsity order of the signal, which can be viewed as model order selection procedure. In this paper, we address the problem of sparsity order estimation in sparse signal recovery. This is of main interest in situations where the signal sparsity is unknown or the signal to be recovered is approximately sparse. It is shown that the proposed method also leads to some kind of signal denoising, where the observations are contaminated with noise. Finally, the performance of the proposed approach is evaluated in different scenarios and compared to an existing method, which shows the effectiveness of the proposed method in terms of order selection as well as denoising.

Keywords: compressed sensing, data denoising, model order selection, sparse representation

Procedia PDF Downloads 439
6924 A Transform Domain Function Controlled VSSLMS Algorithm for Sparse System Identification

Authors: Cemil Turan, Mohammad Shukri Salman

Abstract:

The convergence rate of the least-mean-square (LMS) algorithm deteriorates if the input signal to the filter is correlated. In a system identification problem, this convergence rate can be improved if the signal is white and/or if the system is sparse. We recently proposed a sparse transform domain LMS-type algorithm that uses a variable step-size for a sparse system identification. The proposed algorithm provided high performance even if the input signal is highly correlated. In this work, we investigate the performance of the proposed TD-LMS algorithm for a large number of filter tap which is also a critical issue for standard LMS algorithm. Additionally, the optimum value of the most important parameter is calculated for all experiments. Moreover, the convergence analysis of the proposed algorithm is provided. The performance of the proposed algorithm has been compared to different algorithms in a sparse system identification setting of different sparsity levels and different number of filter taps. Simulations have shown that the proposed algorithm has prominent performance compared to the other algorithms.

Keywords: adaptive filtering, sparse system identification, TD-LMS algorithm, VSSLMS algorithm

Procedia PDF Downloads 315
6923 Development of a Few-View Computed Tomographic Reconstruction Algorithm Using Multi-Directional Total Variation

Authors: Chia Jui Hsieh, Jyh Cheng Chen, Chih Wei Kuo, Ruei Teng Wang, Woei Chyn Chu

Abstract:

Compressed sensing (CS) based computed tomographic (CT) reconstruction algorithm utilizes total variation (TV) to transform CT image into sparse domain and minimizes L1-norm of sparse image for reconstruction. Different from the traditional CS based reconstruction which only calculates x-coordinate and y-coordinate TV to transform CT images into sparse domain, we propose a multi-directional TV to transform tomographic image into sparse domain for low-dose reconstruction. Our method considers all possible directions of TV calculations around a pixel, so the sparse transform for CS based reconstruction is more accurate. In 2D CT reconstruction, we use eight-directional TV to transform CT image into sparse domain. Furthermore, we also use 26-directional TV for 3D reconstruction. This multi-directional sparse transform method makes CS based reconstruction algorithm more powerful to reduce noise and increase image quality. To validate and evaluate the performance of this multi-directional sparse transform method, we use both Shepp-Logan phantom and a head phantom as the targets for reconstruction with the corresponding simulated sparse projection data (angular sampling interval is 5 deg and 6 deg, respectively). From the results, the multi-directional TV method can reconstruct images with relatively less artifacts compared with traditional CS based reconstruction algorithm which only calculates x-coordinate and y-coordinate TV. We also choose RMSE, PSNR, UQI to be the parameters for quantitative analysis. From the results of quantitative analysis, no matter which parameter is calculated, the multi-directional TV method, which we proposed, is better.

Keywords: compressed sensing (CS), low-dose CT reconstruction, total variation (TV), multi-directional gradient operator

Procedia PDF Downloads 219
6922 Supervised/Unsupervised Mahalanobis Algorithm for Improving Performance for Cyberattack Detection over Communications Networks

Authors: Radhika Ranjan Roy

Abstract:

Deployment of machine learning (ML)/deep learning (DL) algorithms for cyberattack detection in operational communications networks (wireless and/or wire-line) is being delayed because of low-performance parameters (e.g., recall, precision, and f₁-score). If datasets become imbalanced, which is the usual case for communications networks, the performance tends to become worse. Complexities in handling reducing dimensions of the feature sets for increasing performance are also a huge problem. Mahalanobis algorithms have been widely applied in scientific research because Mahalanobis distance metric learning is a successful framework. In this paper, we have investigated the Mahalanobis binary classifier algorithm for increasing cyberattack detection performance over communications networks as a proof of concept. We have also found that high-dimensional information in intermediate features that are not utilized as much for classification tasks in ML/DL algorithms are the main contributor to the state-of-the-art of improved performance of the Mahalanobis method, even for imbalanced and sparse datasets. With no feature reduction, MD offers uniform results for precision, recall, and f₁-score for unbalanced and sparse NSL-KDD datasets.

Keywords: Mahalanobis distance, machine learning, deep learning, NS-KDD, local intrinsic dimensionality, chi-square, positive semi-definite, area under the curve

Procedia PDF Downloads 38
6921 Sparse-View CT Reconstruction Based on Nonconvex L1 − L2 Regularizations

Authors: Ali Pour Yazdanpanah, Farideh Foroozandeh Shahraki, Emma Regentova

Abstract:

The reconstruction from sparse-view projections is one of important problems in computed tomography (CT) limited by the availability or feasibility of obtaining of a large number of projections. Traditionally, convex regularizers have been exploited to improve the reconstruction quality in sparse-view CT, and the convex constraint in those problems leads to an easy optimization process. However, convex regularizers often result in a biased approximation and inaccurate reconstruction in CT problems. Here, we present a nonconvex, Lipschitz continuous and non-smooth regularization model. The CT reconstruction is formulated as a nonconvex constrained L1 − L2 minimization problem and solved through a difference of convex algorithm and alternating direction of multiplier method which generates a better result than L0 or L1 regularizers in the CT reconstruction. We compare our method with previously reported high performance methods which use convex regularizers such as TV, wavelet, curvelet, and curvelet+TV (CTV) on the test phantom images. The results show that there are benefits in using the nonconvex regularizer in the sparse-view CT reconstruction.

Keywords: computed tomography, non-convex, sparse-view reconstruction, L1-L2 minimization, difference of convex functions

Procedia PDF Downloads 275
6920 Effects of Computer-Mediated Dictionaries on Reading Comprehension and Vocabulary Acquisition

Authors: Mohamed Amin Mekheimer

Abstract:

This study aimed to investigate the effects of paper-based monolingual, pop-up and type-in electronic dictionaries on improving reading comprehension and incidental vocabulary acquisition and retention in an EFL context. It tapped into how computer-mediated dictionaries may have facilitated/impeded reading comprehension and vocabulary acquisition. Findings showed differential effects produced by the three treatments compared with the control group. Specifically, it revealed that the pop-up dictionary condition had the shortest average vocabulary searching time, vocabulary and text reading time, yet with less than the type-in dictionary group but more than the book dictionary group in terms of frequent dictionary 'look-ups' (p<.0001). In addition, ANOVA analyses also showed that text reading time differed significantly across all four treatments, and so did reading comprehension. Vocabulary acquisition was reported as enhanced in the three treatments rather than in the control group, but still with insignificant differences across the three treatments, yet with more differential effects in favour of the pop-up condition. Data also assert that participants preferred the pop-up e-dictionary more than the type-in and paper-based groups. Explanations of the findings vis-à-vis the cognitive load theory were presented. Pedagogical implications and suggestions for further research were forwarded at the end.

Keywords: computer-mediated dictionaries, type-in dictionaries, pop-up dictionaries, reading comprehension, vocabulary acquisition

Procedia PDF Downloads 396
6919 A New Framework for ECG Signal Modeling and Compression Based on Compressed Sensing Theory

Authors: Siavash Eftekharifar, Tohid Yousefi Rezaii, Mahdi Shamsi

Abstract:

The purpose of this paper is to exploit compressed sensing (CS) method in order to model and compress the electrocardiogram (ECG) signals at a high compression ratio. In order to obtain a sparse representation of the ECG signals, first a suitable basis matrix with Gaussian kernels, which are shown to nicely fit the ECG signals, is constructed. Then the sparse model is extracted by applying some optimization technique. Finally, the CS theory is utilized to obtain a compressed version of the sparse signal. Reconstruction of the ECG signal from the compressed version is also done to prove the reliability of the algorithm. At this stage, a greedy optimization technique is used to reconstruct the ECG signal and the Mean Square Error (MSE) is calculated to evaluate the precision of the proposed compression method.

Keywords: compressed sensing, ECG compression, Gaussian kernel, sparse representation

Procedia PDF Downloads 421