Martin Weyssow
Ph.D. student at DIRO, University of Montreal and GEODES research lab.

2238 Pavillon André-Aisenstadt
2920 Ch de la Tour
Montreal, QC H3T 1N8
I am a Ph.D. student in the Department of computer science and operations research at the University of Montreal advised by Prof. Houari Sahraoui. I am part of GEODES software engineering research lab where I carry out my research projects.
My research is all about the application of deep learning and natural language processing in software engineering. As part of my current research, my objective is to design data-intensive approaches that can be leveraged to ease the development and maintenance of complex software systems by assisting developers.
The objective of my Ph.D. thesis is twofold. First, it involves the design of multimodal learning approaches to better modeling programs. Secondly, I believe that making deep learning models of code adaptable and robust to adversarial situations where important distribution drift of the data occur is crucial. Therefore, my second objective is to study the resilience of deep learning models of code to these scenarios, and design methods to improve their robustness.
Besides, I enjoy reading about other research topics including psychology, neuroscience, and, deep reinforcement learning. In my free time, I play and compose music (check out this page!).
news
Apr 29, 2022 | I was awarded a prestigious FRQ Merit Scholarship for Foreign Students (PBEEE) funding my Ph.D. |
---|---|
Mar 22, 2022 | I am serving as a PC member of the Deep Learning for Code workshop at ICLR 2022. |
Dec 30, 2021 | Our paper “Better Modeling the Programming World with Code Concept Graphs-augmented Multi-modal Learning” got accepted at ICSE-NIER 2022. |
Dec 17, 2021 | Our paper “Recommmending Metamodel Concepts during Mdoeling Activities with Pre-Trained Language Models” got accepted for SoSym journal AI4MDE theme issue. |
May 1, 2021 |
I am recipient of a Google Scholarship for Excellence offered by DIRO, University of Montreal funding 1-year of full-time research. ![]() ![]() |
selected publications
-
ArXivAST-Probe: Recovering abstract syntax trees from hidden representations of pre-trained language modelsarXiv preprint arXiv:2206.11719 2022
-
ICSE-NIERBetter Modeling the Programming World with Code Concept Graphs-augmented Multi-modal LearningIn 2022 IEEE/ACM 44th International Conference on Software Engineering: New Ideas and Emerging Results (ICSE-NIER) 2022
-
SoSymRecommending Metamodel Concepts during Modeling Activities with Pre-Trained Language ModelsSoftware and Systems Modeling 2021
-
SoSym