An important step in speaker verification is extracting features that best characterize the speaker voice. This paper investigates a front-end processing that aims at improving the performance of speaker verification based on the SVMs classifier, in text independent mode. This approach combines features based on conventional Mel-cepstral Coefficients (MFCCs) and Line Spectral Frequencies (LSFs) to constitute robust multivariate feature vectors. To reduce the high dimensionality required for training these feature vectors, we use a dimension reduction method called principal component analysis (PCA). In order to evaluate the robustness of these systems, different noisy environments have been used. The obtained results using TIMIT database showed that, using the paradigm that combines these spectral cues leads to a significant improvement in verification accuracy, especially with PCA reduction for low signal-to-noise ratio noisy environment.

We propose a new analytical approximation to the $\chi^2$ kernel that converges geometrically. The analytical approximation is derived with elementary methods and adapts to the input distribution for optimal convergence rate. Experiments show the new approximation leads to improved performance in image classification and semantic segmentation tasks using a random Fourier feature approximation of the $\exp-\chi^2$ kernel. Besides, out-of-core principal component analysis (PCA) methods are introduced to reduce the dimensionality of the approximation and achieve better performance at the expense of only an additional constant factor to the time complexity. Moreover, when PCA is performed jointly on the training and unlabeled testing data, further performance improvements can be obtained. Experiments conducted on the PASCAL VOC 2010 segmentation and the ImageNet ILSVRC 2010 datasets show statistically significant improvements over alternative approximation methods.

Estimation of facial expressions, as spatio-temporal processes, can take advantage of kernel methods if one considers facial landmark positions and their motion in 3D space. We applied support vector classification with kernels derived from dynamic time-warping similarity measures. We achieved over 99% accuracy - measured by area under ROC curve - using only the 'motion pattern' of the PCA compressed representation of the marker point vector, the so-called shape parameters. Beyond the classification of full motion patterns, several expressions were recognized with over 90% accuracy in as few as 5-6 frames from their onset, about 200 milliseconds.

In domains like bioinformatics, information retrieval and social network analysis, one can find learning tasks where the goal consists of inferring a ranking of objects, conditioned on a particular target object. We present a general kernel framework for learning conditional rankings from various types of relational data, where rankings can be conditioned on unseen data objects. We propose efficient algorithms for conditional ranking by optimizing squared regression and ranking loss functions. We show theoretically, that learning with the ranking loss is likely to generalize better than with the regression loss. Further, we prove that symmetry or reciprocity properties of relations can be efficiently enforced in the learned models. Experiments on synthetic and real-world data illustrate that the proposed methods deliver state-of-the-art performance in terms of predictive power and computational efficiency. Moreover, we also show empirically that incorporating symmetry or reciprocity properties can improve the generalization performance.

We study the problem of learning with label proportions in which the training data is provided in groups and only the proportion of each class in each group is known. We propose a new method called proportion-SVM, or $\propto$SVM, which explicitly models the latent unknown instance labels together with the known group label proportions in a large-margin framework. Unlike the existing works, our approach avoids making restrictive assumptions about the data. The $\propto$SVM model leads to a non-convex integer programming problem. In order to solve it efficiently, we propose two algorithms: one based on simple alternating optimization and the other based on a convex relaxation. Extensive experiments on standard datasets show that $\propto$SVM outperforms the state-of-the-art, especially for larger group sizes.

We propose one-class support measure machines (OCSMMs) for group anomaly detection which aims at recognizing anomalous aggregate behaviors of data points. The OCSMMs generalize well-known one-class support vector machines (OCSVMs) to a space of probability measures. By formulating the problem as quantile estimation on distributions, we can establish an interesting connection to the OCSVMs and variable kernel density estimators (VKDEs) over the input space on which the distributions are defined, bridging the gap between large-margin methods and kernel density estimators. In particular, we show that various types of VKDEs can be considered as solutions to a class of regularization problems studied in this paper. Experiments on Sloan Digital Sky Survey dataset and High Energy Particle Physics dataset demonstrate the benefits of the proposed framework in real-world applications.

In this work, a methodology is developed to detect sentient actors in spoken stories. Meta-tags are then saved to XML files associated with the audio files. A recursive approach is used to find actor candidates and features which are then classified using machine learning approaches. Results of the study indicate that the methodology performed well on a narrative based corpus of children’s stories. Using Support Vector Machines for classification, an F-measure accuracy score of 86% was achieved for both named and unnamed entities. Additionally, feature analysis indicated that speech features were very useful when detecting unnamed actors.

Currently, signature based Intrusion Detection Systems (IDS) approaches are inadequate to address threats posed to networked systems by zero-day exploits. Statistical machine learning techniques offer a great opportunity to mitigate these threats. However, at this point, statistical based IDS systems are not mature enough to be implemented in realtime systems and the techniques to be used are not sufficiently understood. This study focuses on a recently expanded corpus for IDS analysis. Feature analysis and Support Vector Machines classification are performed to obtain a better understanding of the corpus and to establish a baseline set of results which can be used by other studies for comparison. Results of the classification and feature analysis are discussed.

Support vector machines (SVMs) are among the most used methods for pattern recognition. Acceptable results have been obtained with such methods in many domains and applications. However, as most learning algorithms, SVMs have hyperparameters that influence the effectiveness of the generated model. Thus, choosing adequate values for such hyperparameters is critical in order to obtain satisfactory results for a given classification task, a problem known as model selection. This paper introduces a novel model selection approach for SVMs based on multi-objective optimization and on the bias and variance definition. We propose an evolutionary algorithm that aims to select the configuration of hyperparameters that optimizes a trade-off between estimates of bias and variance; two factors that are closely related to the model accuracy and complexity. The proposed technique is evaluated using a suite of benchmark data sets for classification. Experimental results show the validity of our approach. We found that the model selection criteria resulted very helpful for selecting highly effective classification models.

Author identification is a subfield of Natural Language Processing (NLP) that uses machine learning techniques to identify the author of a text. Most previous research focused on long texts with the assumption that a minimum text length threshold exists under which author identification would no longer be effective. This paper examines author identification in short texts far below this threshold, focusing on messages retrieved from Twitter (maximum length: 140 characters) to determine the most effective feature set for author identification. Both Bag-of-Words (BOW) and Style Marker feature sets were extracted and evaluated through a series of 15 experiments involving up to 12 authors with large and small dataset sizes. Support Vector Machines (SVM) were used for all experiments. Our results achieve classification accuracies approaching that of longer texts, even for small dataset sizes of 60 training instances per author. Style Marker feature sets were found to be significantly more useful than BOW feature sets as well as orders of magnitude faster, and are therefore suggested for potential applications in future research.