Novel Approaches for Practical Machine Learning

Machine learning is a subfield of artificial intelligence that aims at producing computing models from observations (data), with no explicit coding made by humans. Recent advances have illustrated a strong potential of machine learning, with the potential of being a disruptive technology in many domains. For the current project, we are investigating techniques for making practical machine learning. Four main axes are considered: 1) to deal with big unstructured datasets, 2) to learn with a diverse set of representations of the data, 3) to learn from streams of data sensed or produced in real-time, and 4) to develop methods allowing fully automated machine learning with little or no insights from human experts. The internships will allow exploring key technologies that would support the development of applications such as smart cameras, wearable personal devices, and black-box machine learning software. It aims at exploring promising concepts with high commercialization potential.

Mahdieh Abbasi, Vincent Poiré, Julien-Charles Lévesque, Audrey Durand, Zahra Rezaei, Sophie Baillargeon, Ahmed Najjar, Marc-André Gardner, Farkhondeh Kiaee, Olivier Gagnon, Karol Lina Lopez
Faculty Supervisor: 
Christian Gagne
Project Year: