DBTB INT Scott Clark r
summary
Claims Bayesian optimization outperforms grid and random search for hyperparameter tuning, reducing experiment time and cost; reports applicability to neural networks and gradient‑boosted trees trained on datasets that can reach petabyte scale; references SigOpt as a hosted service to manage and automate tuning experiments; advises hands‑on practice using public datasets and competitions like Kaggle to develop intuition and produce features that improve model business value.
AI-written summary · about
related talks
18mAlexandra Johnson, Software Engineer, SigOpt, NYC 2017
18mBest Practices for Hyperparameter Optimization
24mUsing Bayesian Optimization to Tune Machine Learning Models
26mScott Clark, CEO & Co-Founder, SigOpt at MLconf Seattle 2017
38mA Guide to Modern Hyperparameters Turning Algorithms
34mBayesian Optimization and it's application to Neural Networks
25mHyperparameter Tuning and Visualization in Deep Learning
28mBayesian Optimization- Fundamentals, Implementation, and Practice