lab

The UWECEML of lab Professor Mark Crowley in the ECE department at the University of Waterloo carries out fundamental and applied research on the broad area Artificial Intelligence and on Deep Learning and Reinforcement Learning in particular.

The focus of the lab is using real-world problems to discover computationally hard questions for modelling uncertainty, learning predictive models and enabling decision-making. Members of the lab carry out research into single-agent and multi-agent Reinforcement Learning large-scale 2D/3D image-like processing, causal inference/learning from data, and manifold learning/dimensionality reduction. Some of this research is motivated by theoretical opportunities but most of the work flows out of challenges raised by real-world domains including: forest fire management, the automotive domain, medical imaging, digital chemistry/material design and even some video games such as MineCraft.

• Lab website: https://markcrowley.ca
• Reading Groups: If you doing research in AI/ML/RL at Waterloo and want to join in to our weekly reading group you can find out more here: https://markcrowley.ca/reading-groups/

ChemGymRL Project

One exciting ongoing project in the AI for Science area is the application of Reinforcement Learning (RL) to Digital Chemistry. Our new ChemGymRL Open Source Simulator enables the use of Reinforcement Learning (RL) algorithms to train agents towards the target of operating individual chemistry benches given specific material targets. The environment can be thought of as a virtual chemistry laboratory consisting of different stations (or benches) where a variety of tasks can be completed. The environment supports the training of RL agents by associating positive and negative rewards based on the procedure and outcomes of actions taken by the agents. The aim is for ChemGymRL to help bridge the gap between autonomous laboratories and digital chemistry. This will have impacts for producing new materials, chemicals, and drugs. It will also require many technologies including search, feedback and control, and optimization, and artificial intelligence algorithms that can deal with the unique challenges of material design. Find out more in our recent preprint paper on the ChemGymRL project.

Driver Behaviour Learning Project

An ongoing project in the lab collection and analysis of high resolution, multi-modal data of human driver behaviour in real world driving scenarios. This Driver Behaviour Learning project was in collaboration with Magna International and data was collected by volunteer participants before, during and after the pandemic, driving around the Waterloo region in a specially modified vehicle with additional sensors.

The resulting data is being used to learn predictive models of driver behaviour including aggressive-driver labelling and causal analysis of pedestrian movement. Bettering such models can then be used to enable the next generation of Advanced Adaptive Driver Assistance Systems (ADAS) such as intelligent cruise-control and eventually autonomous driving.

Highlighted Projects

The lab has worked with industrial partners on topics in Medical Imaging including classification and labelling of medical images from Digital Pathology and brain scans of Alzheimer’s patients.