معرفی
Marco Pedersoli serves as an Assistant Professor at École de technologie supérieure (ETS) in Montreal since February 2017, where he leads research in computer vision and machine learning. His work focuses on reducing computational costs and annotation requirements for deploying vision algorithms on embedded devices, positioning ETS at the forefront of Montreal's AI ecosystem.
His academic journey includes:
- Ph.D. from Autonomous University of Barcelona (UAB) under Jordi Gonzàlez and Juan José Villanueva
- Post-doctoral research at INRIA Grenoble with Cordelia Schmid and Jakob Verbeek (2015-2016)
- Research at KU Leuven with Tinne Tuytelaars (2012-2015)
Dr. Pedersoli's research tackles deep learning bottlenecks throughweakly-supervised methodologies andcomputational efficiency innovations. His three core projects address:
- Reduced Supervision: Developing weakly/semi-supervised learning for images, video, audio and text
- Exploration Learning: Optimizing data selection in unstructured environments
- Efficient Computation: Accelerating deep learning training and inference
Publication trends (2014-2022) reveal consistent focus onweak supervision (60% of works) andcomputational efficiency (30%), with recent expansion into medical imaging and multimodal emotion recognition. Key venues include CVPR, ICCV, NeurIPS and ECCV.
His accolades include:
- Best Paper Award at ICIAR 2019
- NVIDIA Titan X Pascal hardware donation
Dr. Pedersoli actively mentors 18 graduate students across PhD and MSc programs, with notable placements at Huawei and Radio Canada. His lab secures competitive tax-free funding for projects with international collaborations, including Element AI and European institutions. Current openings emphasize Python/C++ proficiency and deep learning expertise.
He leads a dynamic research group at ETS developing open-source tools for Roi-Pooling, weakly-supervised detection, and 3D object recognition, maintaining active GitHub repositories with community contributions. Recent WACV 2023 acceptances demonstrate ongoing productivity following medical leave.
