Julien Epps

Bio: Julien Epps is an academic researcher from University of New South Wales. The author has contributed to research in topics: Speaker recognition & Speech processing. The author has an hindex of 39, co-authored 257 publications receiving 8270 citations. Previous affiliations of Julien Epps include Motorola & NICTA.

Information Theoretic Measures for Clusterings Comparison: Variants, Properties, Normalization and Correction for Chance

Abstract: Information theoretic measures form a fundamental class of measures for comparing clusterings, and have recently received increasing interest. Nevertheless, a number of questions concerning their properties and inter-relationships remain unresolved. In this paper, we perform an organized study of information theoretic measures for clustering comparison, including several existing popular measures in the literature, as well as some newly proposed ones. We discuss and prove their important properties, such as the metric property and the normalization property. We then highlight to the clustering community the importance of correcting information theoretic measures for chance, especially when the data size is small compared to the number of clusters present therein. Of the available information theoretic based measures, we advocate the normalized information distance (NID) as a general measure of choice, for it possesses concurrently several important properties, such as being both a metric and a normalized measure, admitting an exact analytical adjusted-for-chance form, and using the nominal [0,1] range better than other normalized variants.

The Geneva Minimalistic Acoustic Parameter Set (GeMAPS) for Voice Research and Affective Computing

Abstract: Work on voice sciences over recent decades has led to a proliferation of acoustic parameters that are used quite selectively and are not always extracted in a similar fashion. With many independent teams working in different research areas, shared standards become an essential safeguard to ensure compliance with state-of-the-art methods allowing appropriate comparison of results across studies and potential integration and combination of extraction and recognition systems. In this paper we propose a basic standard acoustic parameter set for various areas of automatic voice analysis, such as paralinguistic or clinical speech analysis. In contrast to a large brute-force parameter set, we present a minimalistic set of voice parameters here. These were selected based on a) their potential to index affective physiological changes in voice production, b) their proven value in former studies as well as their automatic extractability, and c) their theoretical significance. The set is intended to provide a common baseline for evaluation of future research and eliminate differences caused by varying parameter sets or even different implementations of the same parameters. Our implementation is publicly available with the openSMILE toolkit. Comparative evaluations of the proposed feature set and large baseline feature sets of INTERSPEECH challenges show a high performance of the proposed set in relation to its size.

Information theoretic measures for clusterings comparison: is a correction for chance necessary?

Abstract: Information theoretic based measures form a fundamental class of similarity measures for comparing clusterings, beside the class of pair-counting based and set-matching based measures. In this paper, we discuss the necessity of correction for chance for information theoretic based measures for clusterings comparison. We observe that the baseline for such measures, i.e. average value between random partitions of a data set, does not take on a constant value, and tends to have larger variation when the ratio between the number of data points and the number of clusters is small. This effect is similar in some other non-information theoretic based measures such as the well-known Rand Index. Assuming a hypergeometric model of randomness, we derive the analytical formula for the expected mutual information value between a pair of clusterings, and then propose the adjusted version for several popular information theoretic based measures. Some examples are given to demonstrate the need and usefulness of the adjusted measures.

Signal Processing in Sequence Analysis: Advances in Eukaryotic Gene Prediction

Abstract: Genomic sequence processing has been an active area of research for the past two decades and has increasingly attracted the attention of digital signal processing researchers in recent years. A challenging open problem in deoxyribonucleic acid (DNA) sequence analysis is maximizing the prediction accuracy of eukaryotic gene locations and thereby protein coding regions. In this paper, DNA symbolic-to-numeric representations are presented and compared with existing techniques in terms of relative accuracy for the gene and exon prediction problem. Novel signal processing-based gene and exon prediction methods are then evaluated together with existing approaches at a nucleotide level using the Burset/Guigo1996, HMR195, and GENSCAN standard genomic datasets. A new technique for the recognition of acceptor splice sites is then proposed, which combines signal processing-based gene and exon prediction methods with an existing data-driven statistical method. By comparison with the acceptor splice site detection method used in the gene-finding program GENSCAN, the proposed DSP-statistical hybrid technique reveals a consistent reduction in false positives at different levels of sensitivity, averaging a 43% reduction when evaluated on the GENSCAN test set.

Pattern Recognition and Machine Learning

Abstract: Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Machine Learning : A Probabilistic Perspective

Abstract: Today's Web-enabled deluge of electronic data calls for automated methods of data analysis. Machine learning provides these, developing methods that can automatically detect patterns in data and then use the uncovered patterns to predict future data. This textbook offers a comprehensive and self-contained introduction to the field of machine learning, based on a unified, probabilistic approach. The coverage combines breadth and depth, offering necessary background material on such topics as probability, optimization, and linear algebra as well as discussion of recent developments in the field, including conditional random fields, L1 regularization, and deep learning. The book is written in an informal, accessible style, complete with pseudo-code for the most important algorithms. All topics are copiously illustrated with color images and worked examples drawn from such application domains as biology, text processing, computer vision, and robotics. Rather than providing a cookbook of different heuristic methods, the book stresses a principled model-based approach, often using the language of graphical models to specify models in a concise and intuitive way. Almost all the models described have been implemented in a MATLAB software package--PMTK (probabilistic modeling toolkit)--that is freely available online. The book is suitable for upper-level undergraduates with an introductory-level college math background and beginning graduate students.

A general species delimitation method with applications to phylogenetic placements

Abstract: Motivation: Sequence-based methods to delimit species are central to DNA taxonomy, microbial community surveys and DNA metabarcoding studies. Current approaches either rely on simple sequence similarity thresholds (OTU-picking) or on complex and compute-intensive evolutionary models. The OTU-picking methods scale well on large datasets, but the results are highly sensitive to the similarity threshold. Coalescent-based species delimitation approaches often rely on Bayesian statistics and Markov Chain Monte Carlo sampling, and can therefore only be applied to small datasets. Results: We introduce the Poisson tree processes (PTP) model to infer putative species boundaries on a given phylogenetic input tree. We also integrate PTP with our evolutionary placement algorithm (EPA-PTP) to count the number of species in phylogenetic placements. We compare our approaches with popular OTU-picking methods and the General Mixed Yule Coalescent (GMYC) model. For de novo species delimitation, the stand-alone PTP model generally outperforms GYMC as well as OTU-picking methods when evolutionary distances between species are small. PTP neither requires an ultrametric input tree nor a sequence similarity threshold as input. In the open reference species delimitation approach, EPA-PTP yields more accurate results than de novo species delimitation methods. Finally, EPA-PTP scales on large datasets because it relies on the parallel implementations of the EPA and RAxML, thereby allowing to delimit species in high-throughput sequencing data. Availability and implementation: The code is freely available at www.

Information Theoretic Measures for Clusterings Comparison: Variants, Properties, Normalization and Correction for Chance

Abstract: Information theoretic measures form a fundamental class of measures for comparing clusterings, and have recently received increasing interest. Nevertheless, a number of questions concerning their properties and inter-relationships remain unresolved. In this paper, we perform an organized study of information theoretic measures for clustering comparison, including several existing popular measures in the literature, as well as some newly proposed ones. We discuss and prove their important properties, such as the metric property and the normalization property. We then highlight to the clustering community the importance of correcting information theoretic measures for chance, especially when the data size is small compared to the number of clusters present therein. Of the available information theoretic based measures, we advocate the normalized information distance (NID) as a general measure of choice, for it possesses concurrently several important properties, such as being both a metric and a normalized measure, admitting an exact analytical adjusted-for-chance form, and using the nominal [0,1] range better than other normalized variants.