Episode 29 | Tengyu Ma

Non-convex Optimization for Machine Learning: Design, Analysis, and Understanding

Papers and Links:

• Tengyu's Homepage
• Google Scholar
• Thesis
• Off the Convex Path: Trajectories
• Off the Convex Path: Mismatches between Traditional Optimization Analyses and Modern Deep Learning
• Regularization Matters: Generalization and Optimization of Neural Nets v.s. their Induced Kernel
• The Implicit and Explicit Regularization Effects of Dropout
• Why Do Pretrained Language Models Help in Downstream Tasks? An Analysis of Head and Prompt Tuning
• Theoretical Analysis of Self-Training with Deep Networks on Unlabeled Data
• Simplifying Models with Unlabeled Output Data

Listen: