Deep Learning and Convex Optimization: A Deep Dive into Optimization of Deep Neural Networks
Optimization in Data Science Project (graduate course)
Here we discussed and analyzed the effect of over-parameterization on optimization of deep neural networks and implemented a simple 2layer over-parameterized neural network on MNIST dataset to demonstrate this phenomenon and the double descent curve in Julia!
The final report was written in persian due to the class rules for project, yet the slides are english for which was our presentation.