Mark Thomas

deep learning for bioacoustics &
AI for social good

mark [dot] thomas @ dal [dot] ca


I am a PhD candidate in computer science @ Dalhousie University under the supervision of Stan Matwin. My main interests are in applying deep learning to the area of bioacoustics. Specifically, I work on algorithms for detecting endangered baleen whale vocalizations in acoustic recordings.

I hold a MSc and BSc from Acadia University. During my MSc, I was supervised by Pritam Ranjan and Holger Teismann and worked on Bayesian optimization for ecology models.

I am also the co-founder of Dartmouth Runners , a free social running club in the heart of the Darkside. Check us out and join us for a run!

News!

Starting in June, I will be joining X (formally Google X) for a four month PhD Residency in AI.


My co-authors and I have recently submitted two journal articles currently under review:

  • M. Thomas, B. Martin, and S. Matwin, (2021) Improving the performance of a passive acoustic monitoring classification system through semi-supervised deep learning. Special Issue on Machine Learning, Journal of the Acoustical Society of America (JASA)

  • A. Theissler, M. Thomas, M. Burch, F. Gerschner (2021) ConfusionVis: Comparative evaluation and selection of multi-class classifiers based on confusion matrices Knowledge-Based Systems

I recently presented my work using semi-supervised learning to leverage unlabeled PAM data at the ICLR 2021 workshop titled: "From Shallow to Deep: Overcoming Limited and Adverse Data".

Recent Publications:

  • M. Thomas, B. Martin, and S. Matwin. (2021) Leveraging unlabelled data through semi-supervised learning to improve the performance of a marine mammal classification system From Shallow to Deep: Overcoming Limited and Adverse Data Workshop at International Confernence on Learning Representations (ICLR).

  • M. Thomas, B. Martin, K. Kowarski, B. Gaudet, and S. Matwin. (2019) Detecting endangered baleen whales within acoustic recordings using region-based convolutional neural networks. Joint Workshop on AI for Social Good at Neural Information Processing Systems (NeurIPS).

Education:

  • PhD: Computer Science, Dalhousie University (exp. 2022)

  • MSc: Mathematics & Statistics, Acadia University

  • BSc: Mathematics & Statistics and Economics, Acadia University

Work Experience:

  • Intern Machine Learning Scientist @ JASCO Applied Sciences (Jan 2018 - Present)

  • Data Scientist @ Fix.com (May 2015 - Aug 2017)