Completed
Portfolio Construction: Why Without Meta Features?
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Evolution of Efficient and Robust AutoML Systems
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Traditional Machine Learning and AutoML Traditional Ml practice
- 3 Focus of This Talk: Progression of Our Ready-To-Use AutoML Systems
- 4 Decades of Work On Predecessors of AutoML European projects
- 5 Auto-WEKA
- 6 Practical Improvement 1: Intermittent Results Retrieval
- 7 Practical Improvement 2: Reduced Search Space
- 8 Multi-Fidelity Optimization: General Overview
- 9 Multi-Fidelity Optimization: Methods
- 10 Multi-Fidelity Optimization: Results
- 11 Portfolio Construction: Why Without Meta Features?
- 12 Portfolio Construction: Problem Definition
- 13 Portfolio Construction: Method
- 14 Portfolio Construction: Theory
- 15 Portfolio Construction: Empirical Results
- 16 Automated policy selection problem Problem. There is no single
- 17 Automated policy selection method
- 18 Automated policy selection: results
- 19 Putting it all together: Auto-sklearn 2.0
- 20 Auto-PyTorch: Multi-Fidelity Optimization
- 21 Auto-PyTorch: Portfolio Construction w/o Meta-Features
- 22 Auto-PyTorch: Comparison Against Previous AutoML Frameworks
- 23 Auto-PyTorch: Evaluation on Image Data
- 24 Regularization Cocktails for SOTA Deep Learning on Tabular Data Combination of 13 different regularizers
- 25 Regularization Cocktails: Evaluation on 40 datasets