• Technical University of Denmark
• Europe
• https://snehadas.github.io/
• 
• https://scholar.google.com/citations?user=eL0_3FUAAAAJ&hl=en
• https://orcid.org/0000-0002-4017-1280
• @dassneh

Human speech processing; Multimodal processing; Robust speech recognition; Speech enhancement; Machine learning; Paralinguistics; Speech-based assistive technology; Natural language processing; Dialogue and discourse; Corpus development; Ethics.

I work on multi-sensory data modelling, that includes speech and audio, biosignals and visual data modalities. My research incorporates shared perspective from statistics, machine-learning, deep-learning and (social) signal processing, with an emphasis on resource constrained setting. I am also interested in questions on ‘AI-alignment’ from legal and human-rights perspectives, besides the ethical and technical aspects. In collaboration with interdisciplinary experts, I am currently researching and developing tools for AI evaluation and impact-assessment within high-risk use-cases of AI with compliance and conformity as end goals. Equity, accessibility, inclusivity are recurrent themes in my research, providing the foundational grounding and direction to my research.