Shokoufandeh, Ali 1965
Overview
Works:  19 works in 36 publications in 1 language and 449 library holdings 

Roles:  Editor, Thesis advisor, Author 
Publication Timeline
.
Most widely held works by
Ali Shokoufandeh
Theoretical aspects of computer science : advanced lectures by
Gholamreza B Khosrovshahi(
)
17 editions published in 2002 in English and held by 424 WorldCat member libraries worldwide
This book presents the revised final versions of eight lectures given by leading researchers at the First Summer School on Theoretical Aspects of Computer Science in Tehran, Iran, in July 2000. The lectures presented are devoted to quantum computation, approximation algorithms, selftesting/correction, algebraic modeling of data, the regularity lemma, multiple access communication and combinatorial designs, graphtheoretical methods in computer vision, and lowdensity paritycheck codes
17 editions published in 2002 in English and held by 424 WorldCat member libraries worldwide
This book presents the revised final versions of eight lectures given by leading researchers at the First Summer School on Theoretical Aspects of Computer Science in Tehran, Iran, in July 2000. The lectures presented are devoted to quantum computation, approximation algorithms, selftesting/correction, algebraic modeling of data, the regularity lemma, multiple access communication and combinatorial designs, graphtheoretical methods in computer vision, and lowdensity paritycheck codes
Manytomany feature matching for structural pattern recognition by Muhammed Fatih Demirci(
Book
)
1 edition published in 2005 in English and held by 2 WorldCat member libraries worldwide
1 edition published in 2005 in English and held by 2 WorldCat member libraries worldwide
Largescale document labeling using supervised sequence embedding by Dmitriy Bespalov(
Book
)
1 edition published in 2012 in English and held by 2 WorldCat member libraries worldwide
A critical component in computational treatment of an automated document labeling is the choice of an appropriate representation. Proper representation captures specific phenomena of interest in data while transforming it to a format appropriate for a classifier. For a text document, a popular choice is the bagofwords (BoW) representation that encodes presence of unique words with nonzero weights such as TFIDF. Extending this model to long, overlapping phrases (ngrams) results in exponential explosion in the dimensionality of the representation. In this work, we develop a model that encodes long phrases in a lowdimensional latent space with a cumulative function of individual words in each phrase. In contrast to BoW, the parameter space of the proposed model grows linearly with the length of the phrase. The proposed model requires only vector additions and multiplications with scalars to compute the latent representation of phrases, which makes it applicable to largescale text labeling problems. Several sentiment classification and binary topic categorization problems will be used to empirically evaluate the proposed representation. The same model can also encode relative spatial distribution of elements in higherdimensional sequences. In order to verify this claim, the proposed model will be evaluated on a largescale image classification dataset, where images are transformed into twodimensional sequences of quantized image descriptors
1 edition published in 2012 in English and held by 2 WorldCat member libraries worldwide
A critical component in computational treatment of an automated document labeling is the choice of an appropriate representation. Proper representation captures specific phenomena of interest in data while transforming it to a format appropriate for a classifier. For a text document, a popular choice is the bagofwords (BoW) representation that encodes presence of unique words with nonzero weights such as TFIDF. Extending this model to long, overlapping phrases (ngrams) results in exponential explosion in the dimensionality of the representation. In this work, we develop a model that encodes long phrases in a lowdimensional latent space with a cumulative function of individual words in each phrase. In contrast to BoW, the parameter space of the proposed model grows linearly with the length of the phrase. The proposed model requires only vector additions and multiplications with scalars to compute the latent representation of phrases, which makes it applicable to largescale text labeling problems. Several sentiment classification and binary topic categorization problems will be used to empirically evaluate the proposed representation. The same model can also encode relative spatial distribution of elements in higherdimensional sequences. In order to verify this claim, the proposed model will be evaluated on a largescale image classification dataset, where images are transformed into twodimensional sequences of quantized image descriptors
Canonical behavior patterns by Walter C Mankowski(
Book
)
1 edition published in 2012 in English and held by 2 WorldCat member libraries worldwide
A common problem in many areas of behavioral research is the analysis of the large volume of protocol data recorded during the execution of tasks. This dissertation describes a new automated method of protocol analysis to nd canonical behaviors a small subset of behavior protocols that are most representative of the full data set. The method I have developed takes advantage of recent algorithmic developments in pattern recognition. By adapting these methods to the analysis of behavior protocols, I provide a new tool for analysts working with large datasets that are infeasible to study using current methods. The method I propose can also be used as an important complement to existing sequential protocol analysis techniques, by allowing researchers to build their models based on a few highly representative samples. The contributions of this dissertation include the adaptation of the method to the analysis of behavior protocols; the development of similarity measures appropriate to behavior protocols; an extension of the method to work in oriented topologies; and a demonstration of the method's utility in realworld problem domains, particularly web browsing and driving
1 edition published in 2012 in English and held by 2 WorldCat member libraries worldwide
A common problem in many areas of behavioral research is the analysis of the large volume of protocol data recorded during the execution of tasks. This dissertation describes a new automated method of protocol analysis to nd canonical behaviors a small subset of behavior protocols that are most representative of the full data set. The method I have developed takes advantage of recent algorithmic developments in pattern recognition. By adapting these methods to the analysis of behavior protocols, I provide a new tool for analysts working with large datasets that are infeasible to study using current methods. The method I propose can also be used as an important complement to existing sequential protocol analysis techniques, by allowing researchers to build their models based on a few highly representative samples. The contributions of this dissertation include the adaptation of the method to the analysis of behavior protocols; the development of similarity measures appropriate to behavior protocols; an extension of the method to work in oriented topologies; and a demonstration of the method's utility in realworld problem domains, particularly web browsing and driving
Graph theoretical methods in object recognition and related problems in extremal graph theory by
Ali Shokoufandeh(
)
1 edition published in 1999 in English and held by 2 WorldCat member libraries worldwide
1 edition published in 1999 in English and held by 2 WorldCat member libraries worldwide
Drexel Object Occlusion Repository (DOOR)(
Book
)
2 editions published in 2005 in English and held by 2 WorldCat member libraries worldwide
The Drexel Object Occlusion Repository is a reference set of images for computer vision and object recognition researchers. The images are constructed by overlapping input objects from the COIL20 database and occluding them by various amounts. The amount of occlusion for each image is measured at the pixel level. An accompanying text file for each occlusion image describes the input files, occlusion rates etc
2 editions published in 2005 in English and held by 2 WorldCat member libraries worldwide
The Drexel Object Occlusion Repository is a reference set of images for computer vision and object recognition researchers. The images are constructed by overlapping input objects from the COIL20 database and occluding them by various amounts. The amount of occlusion for each image is measured at the pixel level. An accompanying text file for each occlusion image describes the input files, occlusion rates etc
Subset selection using nonlinear optimization by Trip Denton(
Book
)
1 edition published in 2007 in English and held by 2 WorldCat member libraries worldwide
1 edition published in 2007 in English and held by 2 WorldCat member libraries worldwide
Problems in extremal and combinatorial geometry by Thomas A Plick(
Book
)
1 edition published in 2012 in English and held by 2 WorldCat member libraries worldwide
This thesis deals with three families of optimization problems: (1) Euclidean optimization problems on random point sets; (2) independent sets in hypergraphs; and (3) packings in point lattices. First, we consider bounds on several monochromatic and bichromatic optimization problems including minimum matching, minimum spanning trees, and the travelling salesman problem. Many of these problems lend themselves to representations in terms of hierarchically separated treestrees with uniform branching factor and depth, and having edge weights exponential in the depth of the edge in the tree. In the second part, we consider the independent set problem on uniform hypergraphs, in anticipation of applying it to the third part, packing problems on point lattices. In these problems we wish to select a subset of points from an n n ::: n grid avoiding particular patterns. We also study several generalizations of these problems that have not been handled previously
1 edition published in 2012 in English and held by 2 WorldCat member libraries worldwide
This thesis deals with three families of optimization problems: (1) Euclidean optimization problems on random point sets; (2) independent sets in hypergraphs; and (3) packings in point lattices. First, we consider bounds on several monochromatic and bichromatic optimization problems including minimum matching, minimum spanning trees, and the travelling salesman problem. Many of these problems lend themselves to representations in terms of hierarchically separated treestrees with uniform branching factor and depth, and having edge weights exponential in the depth of the edge in the tree. In the second part, we consider the independent set problem on uniform hypergraphs, in anticipation of applying it to the third part, packing problems on point lattices. In these problems we wish to select a subset of points from an n n ::: n grid avoiding particular patterns. We also study several generalizations of these problems that have not been handled previously
On the Applications of Metric Trees and Metric Labeling to Hard Combinatorial Optimization Problems by Yusuf Osmanlioglu(
Book
)
1 edition published in 2016 in English and held by 1 WorldCat member library worldwide
Matching of metric distributions is a fundamental problem in computer science having numerous real life applications including computer vision, pattern recognition, and natural language processing. The problem can be reduced to graph matching since arbitrary metrics can be represented using graphs. Exact graph matching is known to be computationally intractable which motivates inexact matching approaches. Although graphs structurally represent metric data without losing information, processing data over graphs is prone to entail performance problems. Specifically, running time of many graph algorithms depend on the number of edges and vertices of the input graph which increases quadratically by the number of nodes on the extreme case of complete graphs. Thus, it is desirable to obtain a sparse representation of the data while preserving the quality of information. A common technique to achieve this is through representing graphs by metric trees which recently became defacto metric structures for embedding problems. In this dissertation, we focus on problems involving data that can be represented by graphs. As a general theme, we concentrate on approximation of graphs by trees for improving the performance of certain algorithms using topological structure of trees. We also focus on the fundamental problem of inexact graph matching, efficient approximation algorithms for the problem and its applications. Specifically, we present an inexact graph matching problem referred to as multilayer matching, which utilizes the structure of hierarchical metric trees. We represent graphs as trees and achieve matching over the trees to improve the performance and accuracy of inexact matching. We also establish a relationship between the well known metric labeling problem and inexact graph matching. We propose efficient approximation algorithms for both multilayer matching and metric labeling using the primaldual approximation scheme. We provide application of the proposed methods to image matching, pattern recognition, and question answering problems. Finally, we present a novel motion segmentation method utilizing metric trees which provides tracking of objects in a video sequence without a priori knowledge of number of objects in the scene
1 edition published in 2016 in English and held by 1 WorldCat member library worldwide
Matching of metric distributions is a fundamental problem in computer science having numerous real life applications including computer vision, pattern recognition, and natural language processing. The problem can be reduced to graph matching since arbitrary metrics can be represented using graphs. Exact graph matching is known to be computationally intractable which motivates inexact matching approaches. Although graphs structurally represent metric data without losing information, processing data over graphs is prone to entail performance problems. Specifically, running time of many graph algorithms depend on the number of edges and vertices of the input graph which increases quadratically by the number of nodes on the extreme case of complete graphs. Thus, it is desirable to obtain a sparse representation of the data while preserving the quality of information. A common technique to achieve this is through representing graphs by metric trees which recently became defacto metric structures for embedding problems. In this dissertation, we focus on problems involving data that can be represented by graphs. As a general theme, we concentrate on approximation of graphs by trees for improving the performance of certain algorithms using topological structure of trees. We also focus on the fundamental problem of inexact graph matching, efficient approximation algorithms for the problem and its applications. Specifically, we present an inexact graph matching problem referred to as multilayer matching, which utilizes the structure of hierarchical metric trees. We represent graphs as trees and achieve matching over the trees to improve the performance and accuracy of inexact matching. We also establish a relationship between the well known metric labeling problem and inexact graph matching. We propose efficient approximation algorithms for both multilayer matching and metric labeling using the primaldual approximation scheme. We provide application of the proposed methods to image matching, pattern recognition, and question answering problems. Finally, we present a novel motion segmentation method utilizing metric trees which provides tracking of objects in a video sequence without a priori knowledge of number of objects in the scene
Metric tree weight adjustment and infinite complete binary trees as groups by Craig Schroeder(
Book
)
1 edition published in 2006 in English and held by 1 WorldCat member library worldwide
1 edition published in 2006 in English and held by 1 WorldCat member library worldwide
Predicting Ecommerce Item Popularity Using Image Quality Features by Stephen Zakrewsky(
Book
)
1 edition published in 2016 in English and held by 1 WorldCat member library worldwide
In order to traverse the plethora of items for sale online, searching, ranking, and recommendation systems must be built, and the quality of these systems can make the difference between boom or bust. In all of these methods, being able to distinguish between popular and nonpopular items is very important. Traditionally, these systems have only utilized textual metadata, however, images represent first order information to the shopper, and are composed of a variety of signals that shoppers respond to. In this thesis we look at the problem of predicting item popularity on a popular ecommerce site using image quality features, and show that these features provide complementary information to the textual features in making this prediction
1 edition published in 2016 in English and held by 1 WorldCat member library worldwide
In order to traverse the plethora of items for sale online, searching, ranking, and recommendation systems must be built, and the quality of these systems can make the difference between boom or bust. In all of these methods, being able to distinguish between popular and nonpopular items is very important. Traditionally, these systems have only utilized textual metadata, however, images represent first order information to the shopper, and are composed of a variety of signals that shoppers respond to. In this thesis we look at the problem of predicting item popularity on a popular ecommerce site using image quality features, and show that these features provide complementary information to the textual features in making this prediction
Theoretical aspects of computer science : advanced lectures(
)
1 edition published in 2002 in English and held by 1 WorldCat member library worldwide
1 edition published in 2002 in English and held by 1 WorldCat member library worldwide
Secure Signal Processing and Secure Machine Learning using Fully Homomorphic Encryption by
Thomas M Shortell(
)
1 edition published in 2018 in English and held by 1 WorldCat member library worldwide
This dissertation focuses on the new techniques for secure and private computation for signal processing and machine learning. Specifically, the thesis will focuses on extending Fully Homomorphic Encryption (FHE) technique in a cloud computing set up by running the algorithms while the data is encrypted. The (FHE) comes at a cost of integer space and not the real space needed by signal processing and machine learning algorithms. Solving this problem requires using numerical models that represent real numbers in an integers space including a rational number format and a fixed point binary format. These models allow the computation of signal processing and machine learning algorithms while the data is encrypted. This dissertation includes analysis and implementation of a natural logarithm, BrightnessContrast filter, Fast Fourier Transform, Speeded Up Robust Features, Histogram of Oriented Gradients, and Convolutional Neural Networks. Analyzing these algorithms with the numerical models provide tight upperbounds on numerical error introduced from their use. Each of the implementations provide unique understanding of error propagation in the encrypted domain. Experimental results of each implementation were aligned with the expected error based on the theorems. Despite their algorithmic constructions, each implementation is a step towards more advanced computations for privacy and security in the cloud
1 edition published in 2018 in English and held by 1 WorldCat member library worldwide
This dissertation focuses on the new techniques for secure and private computation for signal processing and machine learning. Specifically, the thesis will focuses on extending Fully Homomorphic Encryption (FHE) technique in a cloud computing set up by running the algorithms while the data is encrypted. The (FHE) comes at a cost of integer space and not the real space needed by signal processing and machine learning algorithms. Solving this problem requires using numerical models that represent real numbers in an integers space including a rational number format and a fixed point binary format. These models allow the computation of signal processing and machine learning algorithms while the data is encrypted. This dissertation includes analysis and implementation of a natural logarithm, BrightnessContrast filter, Fast Fourier Transform, Speeded Up Robust Features, Histogram of Oriented Gradients, and Convolutional Neural Networks. Analyzing these algorithms with the numerical models provide tight upperbounds on numerical error introduced from their use. Each of the implementations provide unique understanding of error propagation in the encrypted domain. Experimental results of each implementation were aligned with the expected error based on the theorems. Despite their algorithmic constructions, each implementation is a step towards more advanced computations for privacy and security in the cloud
Optimal matching and deterministic sampling by Jeff Abrahamson(
Book
)
1 edition published in 2007 in English and held by 1 WorldCat member library worldwide
1 edition published in 2007 in English and held by 1 WorldCat member library worldwide
Automatic construction, maintenance, and optimization of dynamic agent organizations by Evan Andrew Sultanik(
Book
)
1 edition published in 2010 in English and held by 1 WorldCat member library worldwide
1 edition published in 2010 in English and held by 1 WorldCat member library worldwide
Multimodal Information Retrieval and Classification by Kamelia Aryafar(
Book
)
1 edition published in 2015 in English and held by 1 WorldCat member library worldwide
Classification optimizations are the corner stone of machine learning models. The main goal of classifiers is to utilize all available data modalities in training to boost the classification performance metrics. This thesis deals with classification and retrieval models from two different perspectives: (1) the single modality classification optimizations where only one data channel can be used; (2) the multimodal classification methods where more than one data channel is available; A classification system is composed of two main steps: extraction of meaningful features to represent the dataset in a feature space and the classification optimization. The first contribution of this thesis is based on sparse approximation techniques introduced by Donoho, namely the l1 regression. We introduce a sparsityeager support vector machine optimization that combines the ideas behind l1regression and SVM to boost the classification performance. We show that the optimization of sparsityeager SVM can be relaxed and formulated as a linear program. This linear program is then solved by fast gradient descent techniques, yielding an optimal set of classifier coefficients. We compare the performance of this classifier with stateoftheart deep neural networks and baseline models on various public datasets. The second contribution of this thesis is a vector space model of feature vectors to boost the classification performance. This representation is similar to the explicit semantic analysis modeling of text documents introduced by Evgeniy Gabrilovich and Shaul Markovitch. In essence, the explicit semantic analysis representation is an extension of term frequency inverse document frequency modeling to multidimensional feature vectors. Through a set of experiments, we show that this representation can boost the classification accuracy. The explicit semantic analysis can also provide an efficient crossdomain information retrieval framework. We combine this representation with a canonical correlation optimization to achieve this. The third contribution of this thesis is based on multimodal approaches to classification: sparse linear integration model and l1SVM. We extend the sparsityeager support vector machine optimization to deal with more than one data modality. Again we formulate this optimization as a linear combination of the training samples in multimodal settings. We then relax this optimization by replacing the l0norm. The final optimization is solvable using convex optimization methods. We show that combining all available data channels can boost the classification accuracy of the sparsityeager support vector machine classifier in comparison with baseline classifiers
1 edition published in 2015 in English and held by 1 WorldCat member library worldwide
Classification optimizations are the corner stone of machine learning models. The main goal of classifiers is to utilize all available data modalities in training to boost the classification performance metrics. This thesis deals with classification and retrieval models from two different perspectives: (1) the single modality classification optimizations where only one data channel can be used; (2) the multimodal classification methods where more than one data channel is available; A classification system is composed of two main steps: extraction of meaningful features to represent the dataset in a feature space and the classification optimization. The first contribution of this thesis is based on sparse approximation techniques introduced by Donoho, namely the l1 regression. We introduce a sparsityeager support vector machine optimization that combines the ideas behind l1regression and SVM to boost the classification performance. We show that the optimization of sparsityeager SVM can be relaxed and formulated as a linear program. This linear program is then solved by fast gradient descent techniques, yielding an optimal set of classifier coefficients. We compare the performance of this classifier with stateoftheart deep neural networks and baseline models on various public datasets. The second contribution of this thesis is a vector space model of feature vectors to boost the classification performance. This representation is similar to the explicit semantic analysis modeling of text documents introduced by Evgeniy Gabrilovich and Shaul Markovitch. In essence, the explicit semantic analysis representation is an extension of term frequency inverse document frequency modeling to multidimensional feature vectors. Through a set of experiments, we show that this representation can boost the classification accuracy. The explicit semantic analysis can also provide an efficient crossdomain information retrieval framework. We combine this representation with a canonical correlation optimization to achieve this. The third contribution of this thesis is based on multimodal approaches to classification: sparse linear integration model and l1SVM. We extend the sparsityeager support vector machine optimization to deal with more than one data modality. Again we formulate this optimization as a linear combination of the training samples in multimodal settings. We then relax this optimization by replacing the l0norm. The final optimization is solvable using convex optimization methods. We show that combining all available data channels can boost the classification accuracy of the sparsityeager support vector machine classifier in comparison with baseline classifiers
A unit cell based multiscale modeling and design approach for tissue engineered scaffolds by Connie Gomez(
Book
)
1 edition published in 2007 in English and held by 1 WorldCat member library worldwide
1 edition published in 2007 in English and held by 1 WorldCat member library worldwide
Finding groups of graphs in databases by Mitchell A Peabody(
Book
)
1 edition published in 2002 in English and held by 1 WorldCat member library worldwide
1 edition published in 2002 in English and held by 1 WorldCat member library worldwide
Between a rock and a hard place : Euclidean TSP in the presence of polygonal obstacles by Jeff Abrahamson(
Book
)
1 edition published in 2005 in English and held by 1 WorldCat member library worldwide
1 edition published in 2005 in English and held by 1 WorldCat member library worldwide
more
fewer
Audience Level
0 

1  
Kids  General  Special 
Related Identities
 Khosrovshahi, Gholamreza B. 1939 Author Editor
 Shokrollahi, Amin 1964 Editor
 Denton, Trip Author
 Salvucci, Dario
 Regli, William C.
 Mankowski, Walter C. Author
 Demirci, Muhammed Fatih Author
 Bespalov, Dmitriy Author
 Novatnack, John
 Plick, Thomas A. Author
Associated Subjects
Algorithms Approximation theoryData processing Artificial intelligence Biomedical engineering Browsers (Computer programs) Classification Combinatorial optimization Computational complexity Computer algorithms Computer graphics Computer science Computer software Computer vision Data structures (Computer science) Decision trees Electronic commerce Electronic data processing Engineering models Euclidean algorithm Graph algorithms Graphic methods Graph theory Image processingDigital techniques Information retrieval Lattice theory Mathematical optimization Mathematics Mechanical engineering Motor vehicle driving Polygons Software measurement Text processing (Computer science) Tissue engineering