Evaluating Machine Learning Claims
When and Where
Speakers
Description
Machine learning has already changed many areas of our society, and there is the promise of it revolutionizing others. Why then is there sometimes a difference in the promises made and the reality found in terms of machine learning implementation and efficacy? We will explore how to separate hype from reality in evolving machine learning claims, including discussing what the point of machine learning is, the golden rule of machine learning evaluation, parameters versus hyperparameters, interpolation and extrapolation, and the advantages and disadvantages of statistics for evaluation such as the receiver operating characteristic (ROC) curve and the precision-recall (PR) curve.
About Michael Hoffman
Michael Hoffman creates predictive computational models to understand interactions between genome, epigenome, and phenotype in human cancers. He implemented the genome annotation method Segway, which simplifies interpretation of large multivariate genomic datasets, and was a linchpin of the NIH ENCODE Project analysis. He is a principal investigator at the Princess Margaret Cancer Centre and Associate Professor in the Departments of Medical Biophysics and Computer Science, University of Toronto. He was named a CIHR New Investigator and has received several awards for his academic work, including the NIH K99/R00 Pathway to Independence Award, and the Ontario Early Researcher Award.