Covers regression diagnostics for linear models, emphasizing the importance of checking assumptions and identifying outliers and influential observations.
Explores the provable benefits of overparameterization in model compression, emphasizing the efficiency of deep neural networks and the importance of retraining for improved performance.
