Andreas Nautsch, former member of the da/sec research group, successfully defended his dissertation “Speaker Recognition in Unconstrained Environments” at TU Darmstadt on the 10th of October.
First examiner: Prof. Dr. rer. nat. Max Mühlhäuser
Second examiner: Prof. Dr.-Ing. Christoph Busch
Second examiner: Prof. Dr. Didier Meuwly
Comittee: Prof. Dr. phil. nat. Marc Fischlin, Prof. Dr. Carsten Binnig and Prof. Dr. Arjan Kuijper
His dissertation is concerned with speaker recognition systems in unconstrained environments, which have not been sufficiently researched in terms of making better decisions. Aside from decision making, unconstrained environments imply two other subjects: security and privacy. Within the scope of his dissertation, these research subjects are regarded as both security against short-term replay attacks and privacy preservation within state-of-the-art biometric voice comparators in the light of a potential leak of biometric data. The aforementioned research subjects are united in this dissertation to sustain good decision making processes facing uncertainty from varying signal quality and to strengthen security as well as preserve privacy. Andreas’s contributions to theory, methods and measures made likelihood ratios ready for real world applications with dual use scenarios in forensic sciences (lab validation) and commercial services (application adaptive risk based decision policies), where either consider the biometric system as a white box which needs to compensate for unconstrained environments.
Andreas was involved in many projects and published a considerable number of publications at journals and flagship conferences during his time at da/sec. He acquired two projects (named BioMobile), which were awarded with distinctions by the local state project sponsor; the project consortia brought together industrial and academic partners from speech, identity management, penetration testing and legal fields as well as the German association of banks (BdB) and the German Federal Criminal Police Office (BKA) as associated project partners. He lead our team in a 15+ sites international research team in the 2016 NIST speaker recognition evaluation; was project editor for the ISO/IEC standard on voice data interchange format, and lead a 20+ author inter-disciplinary survey with experts from speech technology, studies of the law & ethics, cryptography and biometrics (the most downloaded article of the Computer Speech & Language journal at the day of his defence). A direct outcome of his doctoral studies at da/sec was the formation of a new special interest group within the International Speech Communication Association (ISCA) on Security & Privacy in Speech Communication.
We are pleased to congratulate Dr. rer. nat. Andreas Nautsch on his promotion and wish him all the best for the future.