Publications & Resources

Conferences

• Disentanglement in a GAN for Unconditional Speech Synthesis
  M. Baas, H. Kamper. In IEEE/ACM TASLP, 2024. [arXiv, code and model checkpoints]
• Voice Conversion for Stuttered Speech, Instruments, Unseen Languages and Textually Described Voices
  M. Baas, H. Kamper. In Proceedings of SACAIR, 2023. [arXiv], [demo page and model]
• Voice Conversion With Just Nearest Neighbors
  M. Baas, B. van Niekerk, H. Kamper. In Proceedings of Interspeech, 2023. [arXiv, code and model checkpoints]
• TransFusion: Transcribing Speech with Multinomial Diffusion
  M. Baas, K. Eloff, H. Kamper. In Proceedings of SACAIR, 2022. [arXiv, code and model checkpoints]
• GAN You Hear Me? Reclaiming Unconditional Speech Synthesis from Diffusion Models
  M. Baas, H. Kamper. In Proceedings of the IEEE SLT Workshop, 2022. [arXiv, audio demo, code and model checkpoints]
• Voice Conversion Can Improve ASR in Very Low-Resource Settings
  M. Baas, H. Kamper. In Proceedings of Interspeech, 2022. [arXiv, audio sample website]
• A Comparison of Discrete and Soft Speech Units for Improved Voice Conversion
  B. van Niekerk, M.A. Carbonneau, J. Zaïdi, M. Baas, H. Seuté, H. Kamper. In Proceedings of IEEE ICASSP, 2022. [arXiv]
• Analyzing Speaker Information in Self-Supervised Models to Improve Zero-Resource Speech Processing
  B. van Niekerk, L. Nortje, M. Baas, and H. Kamper. In Proceedings of Interspeech, 2021. [arXiv]
• StarGAN-ZSVC: Towards Zero-Shot Voice Conversion in Low-Resource Contexts
  M. Baas, H. Kamper. In Proceedings of the Southern African Conference for AI Research (SACAIR), 2020. [bibtex, arXiv, demo website and pretrained models]
• [Student poster] Genetic Algorithms and Competitive Self-Play in Training Deep Neural Networks Agents
  M. Baas. At Deep Learning Indaba poster session, Stellenbosch, 2019.

Competitions

Competition results published at conferences are also included here.

• Winning artwork for Zindi AI in Art Competition
  M. Baas. AI Expo Africa, 2019. [Zindi announcement, AI Expo Africa links (1, 2)]
• 3rd place in Zindi Farm Pin Crop Detection Challence
  M. Baas. Zindi Competition Platform, 2019.
• 2nd place in Uber SANRAL Cape Town Hackathon
  K. Eloff, M. Baas, R. Seebregts. At the Stellenbosch LaunchLab Data Science Hackathon with Zindi, 2019.
• 2nd place in Zindi Yassir ETA prediction Challenge
  M. Baas, K. Eloff, G. Frost, Zindi Competition Platform, 2020.
• 3rd place in Zindi: Automatic Speech Recognition in WOLOF Challenge
  G. Frost, M. Baas, K. Eloff. Zindi Competition Platform, 2021.
• 1st place in LifeQ Cardiac Arrhythmias Prediction Challenge
  M. Baas, Zindi Competition Platform, 2021.
• 1st place in CapeAI Autonomous Shopper Prediction Challenge
  G. Frost, K. Eloff, M. Baas. IndabaX South Africa, 2021.

Invited talks

• Voice conversion with just* nearest neighbors
  IndabaX South Africa, University of Cape Town, 2023.
• Style transfer: Visual art with deep learning
  Maties Machine Learning, Stellenbosch University, 2019.
• Improving implicit speech disentanglement with GANs
  [virtual] Division of Speech, Music and Hearing - KTH Royal Institute of Technology, 2022.

Miscellaneous

• Comparison and Development of Practical Voice Conversion Models
  M. Baas. Skripsie (final undergraduate project) thesis, Stellenbosch University, 2020.
• Disentangled Representations in Speech Processing Applications
  M. Baas. PhD thesis, Stellenbosch University, 2024.

Resources

• Links of useful South African technical product suppliers