Talk 1: Hyperparameter Optimization in Python (40 min + 10 min Q&A)
Speaker: Jakub Czakon
Abstract: Hyperparameter optimization (HPO) is quickly becoming common knowledge. But which library should you use? In this talk, I will review some main HPO libraries including Hyperopt, Scikit-Optimize, Optuna, and HpBandSter. I will present what the algorithms used are doing, and compare the libraries based on the API, speed, visualization utilities, results and more.
Bio: Jakub is a Senior Data Scientist at neptune.ml. He graduated in Physics from the University of Silesia in Katowice, and in Finance from the University of Economics in Wroclaw. He's worked on data science for facial recognition, OCR, cancer detection and classification, satellite image segmentation, text mining labor market data, and more. He was a member of the teams that won MICCAI Munich 2015 "Combined Imaging and Digital Pathology Classification Challenge", won MICCAI Athens 2016 "Pet segmentation challenge using a data management and processing infrastructure", and won crowdAI "Mapping Challenge" competition in 2018.
Talk 2: TBD