Sahidullah

Bio: Sahidullah is an academic researcher from University of Lorraine. The author has contributed to research in topics: Spoofing attack & Speaker recognition. The author has an hindex of 23, co-authored 80 publications receiving 1589 citations. Previous affiliations of Sahidullah include French Institute for Research in Computer Science and Automation & Indian Institute of Technology Kharagpur.

ASVspoof 2019: Future horizons in spoofed and fake audio detection

Abstract: ASVspoof, now in its third edition, is a series of community-led challenges which promote the development of countermeasures to protect automatic speaker verification (ASV) from the threat of spoofing. Advances in the 2019 edition include: (i) a consideration of both logical access (LA) and physical access (PA) scenarios and the three major forms of spoofing attack, namely synthetic, converted and replayed speech; (ii) spoofing attacks generated with state-of-the-art neu-ral acoustic and waveform models; (iii) an improved, controlled simulation of replay attacks; (iv) use of the tandem detection cost function (t-DCF) that reflects the impact of both spoofing and countermeasures upon ASV reliability. Even if ASV remains the core focus, in retaining the equal error rate (EER) as a secondary metric, ASVspoof also embraces the growing importance of fake audio detection. ASVspoof 2019 attracted the participation of 63 research teams, with more than half of these reporting systems that improve upon the performance of two baseline spoofing countermeasures. This paper describes the 2019 database, protocols and challenge results. It also outlines major findings which demonstrate the real progress made in protecting against the threat of spoofing and fake audio.

ASVspoof 2017 Version 2.0: meta-data analysis and baseline enhancements

Abstract: The now-acknowledged vulnerabilities of automatic speaker verification (ASV) technology to spoofing attacks have spawned interests to develop so-called spoofing countermeasures. By providing common databases, protocols and metrics for their assessment, the ASVspoof initiative was born to spear-head research in this area. The first competitive ASVspoof challenge held in 2015 focused on the assessment of countermeasures to protect ASV technology from voice conversion and speech synthesis spoofing attacks. The second challenge switched focus to the consideration of replay spoofing attacks and countermeasures. This paper describes Version 2.0 of the ASVspoof 2017 database which was released to correct data anomalies detected post-evaluation. The paper contains as-yet unpublished meta-data which describes recording and playback devices and acoustic environments. These support the analysis of replay detection performance and limits. Also described are new results for the official ASVspoof baseline system which is based upon a constant Q cesptral coefficient frontend and a Gaussian mixture model backend. Reported are enhancements to the baseline system in the form of log-energy coefficients and cepstral mean and variance normalisation in addition to an alternative i-vector backend. The best results correspond to a 48% relative reduction in equal error rate when compared to the original baseline system.

t-DCF: a Detection Cost Function for the Tandem Assessment of Spoofing Countermeasures and Automatic Speaker Verification

Abstract: The ASVspoof challenge series was born to spearhead research in anti-spoofing for automatic speaker verification (ASV). The two challenge editions in 2015 and 2017 involved the assessment of spoofing countermeasures (CMs) in isolation from ASV using an equal error rate (EER) metric. While a strategic approach to assessment at the time, it has certain shortcomings. First, the CM EER is not necessarily a reliable predic-tor of performance when ASV and CMs are combined. Second, the EER operating point is ill-suited to user authentication applications , e.g. telephone banking, characterised by a high target user prior but a low spoofing attack prior. We aim to migrate from CM-to ASV-centric assessment with the aid of a new tandem detection cost function (t-DCF) metric. It extends the conventional DCF used in ASV research to scenarios involving spoofing attacks. The t-DCF metric has 6 parameters: (i) false alarm and miss costs for both systems, and (ii) prior probabilities of target and spoof trials (with an implied third, nontar-get prior). The study is intended to serve as a self-contained, tutorial-like presentation. We analyse with the t-DCF a selection of top-performing CM submissions to the 2015 and 2017 editions of ASVspoof, with a focus on the spoofing attack prior. Whereas there is little to choose between countermeasure systems for lower priors, system rankings derived with the EER and t-DCF show differences for higher priors. We observe some ranking changes. Findings support the adoption of the DCF-based metric into the roadmap for future ASVspoof challenges, and possibly for other biometric anti-spoofing evaluations.

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.

Analyzing linguistic data: a practical introduction to statistics using R

Abstract: The complete title of this book runs ‘Analyzing Linguistic Data: A Practical Introduction to Statistics using R’ and as such it very well reflects the purpose and spirit of the book. The author guides the reader in about 350 pages from descriptive and basic statistical methods over classification and clustering to (generalised) linear and mixed models. Each of the methods is introduced in the context of concrete linguistic problems and demonstrated on exciting datasets from current research in the language sciences. In line with its practical orientation, the book focuses primarily on using the methods and interpreting the results. This implies that the mathematical treatment of the techniques is held at a minimum if not absent from the book. In return, the reader is provided with very detailed explanations on how to conduct the analyses using R [1]. The first chapter sets the tone being a 20-page introduction to R. For this and all subsequent chapters, the R code is intertwined with the chapter text and the datasets and functions used are conveniently packaged in the languageR package that is available on the Comprehensive R Archive Network (CRAN). With this approach, the author has done an excellent job in enabling researchers and students alike to reproduce the analyses and learn by doing. Another quality as a textbook is the fact that every chapter ends with Workbook sections where the user is invited to exercise his or her analysis skills on supplemental datasets. Full solutions including code, results and comments are given in Appendix A (30 pages). Instructors are therefore very well served by this text, although they might want to balance the book with some more mathematical treatment depending on the target audience. After the introductory chapter on R, the book opens on graphical data exploration. Chapter 3 treats probability distributions and common sampling distributions. Under basic statistical methods (Chapter 4), distribution tests and tests on means and variances are covered. Chapter 5 deals with clustering and classification. Strangely enough, the clustering section has material on PCA, factor analysis, correspondence analysis and includes only one subsection on clustering, devoted notably to hierarchical partitioning methods. The classification part deals with decision trees, discriminant analysis and support vector machines. The regression chapter (Chapter 6) treats linear models, generalised linear models, piecewise linear models and a substantial section on models for lexical richness. The final chapter on mixed models is particularly interesting as it is one of the few text book accounts that introduce the reader to using the (innovative) lme4 package of Douglas Bates which implements linear mixed-effects models. Moreover, the case studies included in this

The ASVspoof 2017 Challenge: Assessing the Limits of Replay Spoofing Attack Detection

Abstract: The ASVspoof initiative was created to promote the development of countermeasures which aim to protect automatic speaker verification (ASV) from spoofing attacks. The first community-led, common evaluation held in 2015 focused on countermeasures for speech synthesis and voice conversion spoofing attacks. Arguably, however, it is replay attacks which pose the greatest threat. Such attacks involve the replay of recordings collected from enrolled speakers in order to provoke false alarms and can be mounted with greater ease using everyday consumer devices. ASVspoof 2017, the second in the series, hence focused on the development of replay attack countermeasures. This paper describes the database, protocols and initial findings. The evaluation entailed highly heterogeneous acoustic recording and replay conditions which increased the equal error rate (EER) of a baseline ASV system from 1.76% to 31.46%. Submissions were received from 49 research teams, 20 of which improved upon a baseline replay spoofing detector EER of 24.77%, in terms of replay/non-replay discrimination. While largely successful, the evaluation indicates that the quest for countermeasures which are resilient in the face of variable replay attacks remains very much alive.

Likelihood Ratios for Out-of-Distribution Detection

Abstract: Discriminative neural networks offer little or no performance guarantees when deployed on data not generated by the same process as the training distribution. On such out-of-distribution (OOD) inputs, the prediction may not only be erroneous, but confidently so, limiting the safe deployment of classifiers in real-world applications. One such challenging application is bacteria identification based on genomic sequences, which holds the promise of early detection of diseases, but requires a model that can output low confidence predictions on OOD genomic sequences from new bacteria that were not present in the training data. We introduce a genomics dataset for OOD detection that allows other researchers to benchmark progress on this important problem. We investigate deep generative model based approaches for OOD detection and observe that the likelihood score is heavily affected by population level background statistics. We propose a likelihood ratio method for deep generative models which effectively corrects for these confounding background statistics. We benchmark the OOD detection performance of the proposed method against existing approaches on the genomics dataset and show that our method achieves state-of-the-art performance. Finally, we demonstrate the generality of the proposed method by showing that it significantly improves OOD detection when applied to deep generative models of images.