This repository contains notes for the "Machine Learning for Design Engineers" module offered at Imperial College London. The module is assessed in three categories: two written reports (15%), lab exam (15%), and final exam (70%). In this repo, you'll find my notes for the course as well as an in-depth analysis of each labs to prepare for the lab exam. You can also find the written reports attached.
Lab 1 - Building an ANN (Artificial Neural Network) with a sigmoid activation function and backpropagation
- Understand and code a simple neuron
- Understand how a neuron learns
- Understand its limitations
Lab 2 - Adding Non-linearity (Hidden layers) to solve XOR problems with softmax activation and cross-entropy loss function
- Understand Backpropagation (Deriving each partial derivatives)
- Write a neural network with 1 or more hidden layers
- Solve the XOR
- Understand how to build general classifiers
Lab 3 - Using PyTorch library to build multiple hidden layers
- Understand how to use PyTorch to write a neural network
- Write a neural network with multiple neural networks
- Write a convolutional network for the MNIST database