Machine learning poster at MMLS 2024

PhD Candidate Matt Raymond presented his work on Joint Optimization of Piecewise Linear Ensembles (JOPLEn) at the 2024 Midwest Machine Learning Symposium in Minneapolis, MN. His poster, titled “Joint Optimization Significantly Improves Gradient Boosting” demonstrates the performance improvements and novel regularization approaches possible using this new approach to learning piecewise linear functions.

a) shows the training data for some function, b) shows the true function, c) shows a competing approach, and d) shows the function learned by JOPLEn.