About me

I am Constance Douwes, a Ph.D. in artificial intelligence, and I study the energy and environmental impact of deep learning models applied to audio signals. Concerned about raising awareness of responsible resource usage, I started my research at IRCAM through neural audio synthesis and developed a methodology to assess the trade-offs between audio quality and the energy efficiency of deep learning models.

Currently, I’m a postdoctoral researcher at Inria, where I continue my research in the field of sound event detection, with the ultimate goal of extending my work on environmental impact to the entire audio domain.