PhD Student in Computer Science or Interdisciplinary Program

PhD student to start in January 2025 or as soon as possible to join a multidisciplinary team with several groups involved including MUN’s Centre for Innovation and Learning in Teaching (CITL), College of the North Atlantic (CAN), and Nova Scotia Community College (NSCC). Student will be working under the supervision of Dr. Gagnon (http://www.ucs.mun.ca/~pgagnon/) and Dr. Peña-Castillo (https://www.cs.mun.ca/~lourdes/).

Project description

Using the large-scale mapping of kelp beds provided by other collaborators, we will establish ground truth (labelled) regions, delineating areas where kelp beds exist (true positives) and where they are absent (true negatives).  These labelled regions, indicating whether kelp beds are present, will be the basis for a self-supervised deep learning approach. This approach will allows us to train deep learning architectures for kelp bed detection using unlabelled satellite images. Beyond the initial  self-supervised training phase, we will implement an active learning  framework. Once an accurate kelp bed detector is generated through this approach, we will apply the classifier to satellite images spanning multiple years to track changes in kelp beds over time, allowing us to  monitor the effects of intervention programs and assess the impact of climate change on kelp bed dynamics. By extending our classifier to temporal datasets, we aim to contribute valuable insights into the dynamics of kelp ecosystems in the face of a warming climate and human interventions.

Additionally, we will develop advanced visualization tools that enhance the interpretation and presentation of satellite remote sensing data. This involves creating interactive, user-friendly interfaces that enable researchers, policymakers, and the public to engage with complex spatial data in a more intuitive and insightful manner. By utilizing technologies such as 3D mapping, augmented reality (AR), and virtual reality (VR), this project aims to transform traditional two-dimensional data representations into dynamic, multi-dimensional visual experiences. This will not only aid in better understanding spatial relationships and patterns in the data but also facilitate more effective communication of findings to a broader audience, thereby making satellite data more accessible and actionable for decision-making processes.

The successful candidate will lead the generation of kelp bed detection models  from Earth Observation (EO) data and advanced visualization tools for satellite data interpretation (see description above), and will work as the liaison between the different groups.

• Master’s degree in Computer Science, Bioinformatics or related discipline.
• Excellent communication skills.
• Knowledge in machine learning and deep learning.
• Strong programming skills in Python and R.
• Ability to work in a multidisciplinary and collaborative research environment.