This repository contains the second homework assignment for the Deep Learning course (Fall 2018) at Sharif University of Technology. The course, instructed by Professor Hamid Beigy, covers fundamental and advanced topics in deep learning, including multi-layer perceptrons, convolutional neural networks, recurrent neural networks, and deep generative models.
The homework focuses on the following key areas:
- Linear Regions in Neural Networks: Analyzing the linear regions identified by neural networks with absolute value activation functions.
- Composition of Functions: Understanding the effects of composing multiple functions on the identification of input regions.
- Layered Neural Networks: Exploring the behavior of neural networks with multiple layers of absolute value activation functions.