DSSS@UofT: Jesse Cresswell
Join us for the next installment of the Data Science Speaker Series at U of T with:
Jesse Cresswell, PhD
Machine Learning Scientist, Layer 6 AI at TD
Free Event | Registration Required
Evaluating Model Performance on Highly Imbalanced Datasets
Imbalanced datasets are extremely common in real-world data science problems, and techniques are widely known for handling imbalance when modelling. After modelling, meaningful evaluation and comparison of models remains a challenging task in its own right. In this seminar I will discuss the ways that dataset imbalance can confound our use of typical model performance metrics, and present challenges for fair comparisons. I will introduce advanced metrics which retain desirable properties even as datasets become skewed, and which give theoretical assurance that cross-model or cross-dataset comparisons are fair. Using these metrics can ensure better model selection in the face of imbalance.gaming.
Jesse is a Machine Learning Scientist at Layer 6 AI within TD, and is the Team Lead for US Credit Risk, TD Insurance, and Quantum Machine Learning. His applied work centers on building machine learning models in high risk and highly regulated domains. Jesse maintains a keen interest in quantum computing and its implications for machine leaning, as well as reinforcement learning.
Before joining Layer 6, Jesse completed his PhD at the Department of Physics, University of Toronto. His research at the intersection of quantum computing and string theory was supported by a Vanier Scholarship.