Sajjad Amini, M. Sadeghi, M. Joneidi, M. Babaie-zadeh, C. Jutten
{"title":"Outlier-aware dictionary learning for sparse representation","authors":"Sajjad Amini, M. Sadeghi, M. Joneidi, M. Babaie-zadeh, C. Jutten","doi":"10.1109/MLSP.2014.6958854","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958854","url":null,"abstract":"Dictionary learning (DL) for sparse representation has been widely investigated during the last decade. A DL algorithm uses a training data set to learn a set of basis functions over which all training signals can be sparsely represented. In practice, training signals may contain a few outlier data, whose structures differ from those of the clean training set. The presence of these unpleasant data may heavily affect the learning performance of a DL algorithm. In this paper we propose a robust-to-outlier formulation of the DL problem. We then present an algorithm for solving the resulting problem. Experimental results on both synthetic data and image denoising demonstrate the promising robustness of our proposed problem.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"194 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123538046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
R. Dürichen, T. Wissel, F. Ernst, Marco A. F. Pimentel, D. Clifton, A. Schweikard
{"title":"A unified approach for respiratory motion prediction and correlation with multi-task Gaussian Processes","authors":"R. Dürichen, T. Wissel, F. Ernst, Marco A. F. Pimentel, D. Clifton, A. Schweikard","doi":"10.1109/MLSP.2014.6958895","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958895","url":null,"abstract":"In extracranial robotic radiotherapy, tumour motion due to respiration is compensated based external markers. Two models are typically used to enable a real-time adaptation. A prediction model, which compensates time latencies of the treatment systems due to e.g. kinematic limitations, and a correlation model, which estimates the internal tumour position based on external markers. We present a novel approach based on multi-task Gaussian Processes (MTGP) which enables an efficient combination of both models by simultaneously learning the correlation and temporal delays between markers. The approach is evaluated using datasets acquired from porcine and human studies. We conclude that the prediction accuracy of MTGP is superior to that of existing methods and can be further increased by using multivariate input data. We investigate the dependency of the number of internal training points and the potential for using the marginal likelihood for model selection.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"317 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122618601","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A new fast Jacobi-like algorithm for non-orthogonal joint diagonalization of real-valued matrices based on a QR parameterization","authors":"V. Maurandi, E. Moreau","doi":"10.1109/MLSP.2014.6958901","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958901","url":null,"abstract":"Non-orthogonal joint diagonalization of a set of real-valued matrices holds a significant place in numerous blind processing issues as independent component analysis and source separation. In this paper, in order to solve this problem, we propose a novel Jacobi-like algorithm based on a QR parameterization. The primary objective of this iterative algorithm is to derive an analytical solution for each two-by-two diagonalizing sub-matrix using a suitable cost function. By computer simulations, we show that the presented algorithm performs well with respect to three other ones from literature including two Jacobi-like algorithms.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122004142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Sparse topic models by parameter sharing","authors":"Hossein Soleimani, David J. Miller","doi":"10.1109/MLSP.2014.6958911","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958911","url":null,"abstract":"We propose a sparse Bayesian topic model, based on parameter sharing, for modeling text corpora. In Latent Dirichlet Allocation (LDA), each topic models all words, even though many words are not topic-specific, i.e. have similar occurrence frequencies across different topics. We propose a sparser approach by introducing a universal shared model, used by each topic to model the subset of words that are not topic-specific. A Bernoulli random variable is associated with each word under every topic, determining whether that word is modeled topic-specifically, with a free parameter, or by the shared model, with a common parameter. Results of our experiments show that our model achieves sparser topic presence in documents and higher test likelihood than LDA.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"112 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116176278","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ryunosuke Hamada, Takatomi Kubo, K. Katahira, Kenta Suzuki, K. Okanoya, K. Ikeda
{"title":"Birdsong analysis using beta process hidden Markov model","authors":"Ryunosuke Hamada, Takatomi Kubo, K. Katahira, Kenta Suzuki, K. Okanoya, K. Ikeda","doi":"10.1109/MLSP.2014.6958848","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958848","url":null,"abstract":"We proposed a new scheme on automatic annotation and analysis for songs of Bengalese finches, that have variability in terms of syllable sequencing. The scheme annotates songs by using the beta process hidden Markov model, a Bayesian non-parametrics method. The annotation was confirmed to agree to the results by the manual annotation by an expert almost perfectly (0.81-1.00) for songs by three out of four Bengalese finches. Our scheme also analyzed the syntactic rules of the birdsongs and found that the difference of the variations in sequencing of four different birds was accounted for by the difference of tutors, giving a new insight to development of syntax.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"296 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114845122","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Non-negative Tensor Factorization with missing data for the modeling of gene expressions in the Human Brain","authors":"S. F. V. Nielsen, Morten Mørup","doi":"10.1109/MLSP.2014.6958919","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958919","url":null,"abstract":"Non-negative Tensor Factorization (NTF) has become a prominent tool for analyzing high dimensional multi-way structured data. In this paper we set out to analyze gene expression across brain regions in multiple subjects based on data from the Allen Human Brain Atlas [1] with more than 40 % data missing in our problem. Our analysis is based on the non-negativity constrained Canonical Polyadic (CP) decomposition where we handle the missing data using marginalization considering three prominent alternating least squares procedures; multiplicative updates, column-wise, and row-wise updating of the component matrices. We examine three gene expression prediction scenarios based on data missing at random, whole genes missing and whole areas missing within a subject. We find that the column-wise updating approach also known as HALS performs the most efficient when fitting the model. We further observe that the non-negativity constrained CP model is able to predict gene expressions better than predicting by the subject average when data is missing at random. When whole genes and whole areas are missing it is in general better to predict by subject averages. However, we find that when whole genes are missing from all subjects the model based predictions are useful. When analyzing the structure of the components derived for one of the best predicting model orders the components identified in general constitute localized regions of the brain. Non-negative tensor factorization based on marginalization thus forms a promising framework for imputing missing values and characterizing gene expression in the human brain. However, care also has to be taken in particular when predicting the genetic expression levels at a whole region of the brain missing as our analysis indicates that this requires a substantial amount of subjects with data for this region in order for the model predictions to be reliable.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125656756","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Joint estimation of multiple undirected graphical models","authors":"Alexandros Georgogiannis, V. Digalakis","doi":"10.1109/MLSP.2014.6958915","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958915","url":null,"abstract":"Gaussian graphical models are of great interest in statistical learning. Since the conditional independence between the variables corresponds to zero entries in the inverse covariance matrix, one can learn the structure of the graph by estimating a sparse inverse covariance matrix from sample data. This is usually done by solving a convex maximum likelihood problem with a l1-regularization term applied on the inverse covariance matrix. In this study, we develop an estimator for such models appropriate for data coming from several datasets that share the same set of variables and a common network substructure. We assume that there exist a few different edges among the networks while the others (edges) are common. To this end, we form an optimization problem that exploits the problem's special structure and we propose an alternating direction method for its solution. We confirm the performance improvement of our method over existing methods in finding the dependence structure on a real dataset.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115104170","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"The 10th annual MLSP competition: First place","authors":"A. Solin, S. Särkkä","doi":"10.1109/MLSP.2014.6958886","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958886","url":null,"abstract":"The goal of the MLSP 2014 Schizophrenia Classification Challenge was to automatically diagnose subjects with schizophrenia based on multimodal features derived from their magnetic resonance imaging (MRI) brain scans. This challenge took place between June 5 and July 20, 2014, and was organized on Kaggle. We present how this classification problem can be solved in terms of a Bayesian machine learning paradigm known as Gaussian process (GP) classification. The proposed solution achieved an AUC score of 0.928, and it ranked first on the Kaggle private leaderboard.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":" 30","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120826716","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A dictionary learning algorithm for sparse coding by the normalized bilateral projections","authors":"Taro Tezuka","doi":"10.1109/MLSP.2014.6958892","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958892","url":null,"abstract":"Sparse coding is a method of expressing the input vector as a linear combination of a few vectors taken from a set of template vectors, often called a dictionary or codebook. A good dictionary is the one that sparse codes most vectors in a given class of possible input vectors. There are currently several proposals to learn a good dictionary from a set of input vectors. Such methods are termed under the title of dictionary learning. We propose a new dictionary learning algorithm, called K-normalized bilateral projections (K-NBP), which is a modification to a widely used dictionary learning method, i.e., K-singular value decomposition (K-SVD). The main idea behind this was to standardize and normalize the input matrix as a preprocessing stage, and to correspondingly normalize the estimated source vectors in the dictionary update stage. The experimental results revealed that our method was fast, and when the number of iterations was limited, it outperformed K-SVD. Also, if only a coarse approximation was needed, it provided results that were almost like those from K-SVD, but with fewer iterations. This indicated that our method was particularly suited to large data sets with many dimensions, where each iteration took a long time.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116809281","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
K. Papachristou, A. Tefas, N. Nikolaidis, I. Pitas
{"title":"Stereoscopic video shot classification based on Weighted Linear Discriminant Analysis","authors":"K. Papachristou, A. Tefas, N. Nikolaidis, I. Pitas","doi":"10.1109/MLSP.2014.6958932","DOIUrl":"https://doi.org/10.1109/MLSP.2014.6958932","url":null,"abstract":"In this paper we propose a framework for stereoscopic video shot classification that includes low-level representations exploiting visual and disparity information and determination of optimal discriminant subspaces based on Linear Discriminant Analysis (LDA). Low-level representations are obtained through various color, disparity and texture descriptors which are applied to shot key frames. A new LDA-based subspace representation is proposed aiming at the optimal utilization of both visual and disparity information. The proposed shot classification framework has been evaluated on football stereoscopic videos providing enhanced classification performance and class discrimination, in comparison to using visual information only and standard LDA.","PeriodicalId":329167,"journal":{"name":"2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129467391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}