Doctoral Researcher in Speech and Language Technology

Aalto University is where science and art meet technology and business. We shape a sustainable future by making research breakthroughs in and across our disciplines, sparking the game changers of tomorrow and creating novel solutions to major global challenges. Our community is made up of 16 000 students and 5 200 employees, including 446 professors. Our campus is in Espoo, Greater Helsinki, Finland.  Diversity is part of who we are, and we actively work to ensure our community’s diversity and inclusiveness. This is why we warmly encourage qualified candidates from all backgrounds to join our community. The research will be carried out at the  Department of Information and Communications Engineering, DICE , at Aalto University, Finland. The project environment offers excellent infrastructure for deep learning, speech, and audio research, including Aalto University’s large-scale scientific computing cluster with CPU and GPU nodes, access to CSC’s national computing infrastructure including LUMI, and the Aalto Acoustics Lab with anechoic chambers, listening rooms, and audio measurement equipment. We are now looking for a Doctoral Researcher in Speech and Language Technology Are you excited about speech and audio AI, deepfake detection, and the question of how we can identify the origin of AI-generated content? We are looking for a  Doctoral Researcher  to join the funded project  COWAMA: Content-based watermarking for AI-generated audio , led by Assistant Professor Lauri Juvela. The rapid development of generative AI has made it possible to create realistic speech and music, but it has also created risks related to misinformation, malicious deepfakes, copyright, ownership, attribution, and accountability. The project addresses these challenges by developing methods for detecting the origin and authenticity of generated and watermarked audio data, improving content-based audio watermarking, and evaluating robustness against removal, spoofing, adversarial attacks, and realistic downstream processing. Your role and goals As the Doctoral Researcher, you will take primary responsibility for detecting generated audio content from multiple sources , and work jointly with the Postdoctoral Researcher on  robustness, evaluation, and generalization . Your work will include: Developing methods for detecting the authenticity and origin of speech and audio content generated by multiple models and containing different watermarking methods. Creating datasets of diverse real and generated speech and audio, including content from different generative models and multiple watermarking methods. Adapting and improving detector baselines for multi-objective detection involving deepfake detection, content origin detection, and watermark identification.  Developing explainable and localized detection methods for watermark and deepfake attribution, especially when generated content is mixed with other audio sources.  Studying robustness against…