Constance Douwes

About Me

I am an associate professor in computer science at Centrale Méditerranée in Marseille (France). I'm specialized in the environmental footprint of modern AI with a particular interest for deep learning audio models.

I did my Ph.D. at IRCAM in Paris (STMS laboratory, ACIDS team) under the supervision of Philippe Esling and Jean-Pierre Briot, where I studied the energy consumption of neural networks applied to music generation.

After that, I worked as a postdoctoral researcher at INRIA in Nancy (LORIA, Multispeech team) under the supervision of Romain Serizel, focusing on the energy footprint of deep learning systems for sound event detection.

I am now a member of the QARMA team at the LIS laboratory, continuing to explore the environmental footprint of AI systems more broadly.

Videos

🎥 Treize minutes marseille (13min) : La fière de l'IA, 2026 [video]

🎥 PhD Defense (~45min) : On the Environmental Impact of Deep Generative Models for Audio, 2023 [video][manuscript]

🎥 Keynote JIM (~1h) : Défis énergétiques et écologiques de l'IA pour la création musicale, 2025 [video][slides]

🎥 GreenDays (~20min): Unraveling the Energy Footprint of Deep Audio Models, 2025 [video][slides]

Publications

📄 C. Douwes, P. Magron, R. Serizel "Exposing the cost of deep learning audio development", ICASSP 2027 (under review). [article][code]

📄 P. Magron, R. Serizel, C. Douwes, "Investigating the Performance and Energy Costs of Replicating Band-Split RNN for Music Source Separation", ICASSP 2027 (under review). [article][code]

📄 C. Douwes, "Défis énergétiques et écologiques de l’IA pour la création musicale", Revue Francophone d'Informatique et Musique, 2025, Frugalité, pérennité et création (11), pp.1-15. [article]

📄 C. Douwes, R. Serizel, "Energy Consumption Trends in Sound Event Detection Systems", ICASSP 2025. [article][code]

📄 C. Douwes, R. Serizel, "From Computation to Consumption: Exploring the Compute-Energy Link for Training and Testing Neural Networks for Sed Systems", DCASE 2024. [article][code]

📄 S. Cornell, J. Ebbers, C. Douwes, et al., "Sound Event Detection with Heterogeneous Data and Missing Labels", DCASE 2024. [article][code]

📄 C. Douwes, R. Serizel, "Normalizing energy consumption for hardware-independent evaluation", MLSP 2024. [article][code]

📄 C. Douwes, G. Bindi, A. Caillon, P. Esling, JP. Briot "Is Quality Enough? Integrating Energy Consumption in a Large-Scale Evaluation of Neural Audio Synthesis Models", ICASSP 2023. [article][code]

📄 C. Douwes, P. Esling, JP. Briot "Energy Consumption of Deep Generative Audio Models", arXiv 2023. [article]

📄 P. Esling, N. Devis, A. Bitton, et al., "Diet Deep Generative Audio Models with Structured Lottery", Proceedings of the 23rd International Conference on Digital Audio Effects (DAFx-20), 2020. [article][code]

Ressources

Slides and TP about the energy and environmental impacts of AI [slides][notebook]

✳ A little DIY origami BD ✳

Done by Shiyata — you can easily download it here.