TDT4171-Methods-in-AI

Exercises for TDT4171-Methods-in-AI at NTNU

Exercise 1 - Uncertainty, Bayesian networks

Exercise 1 implements a Bayesian network using Inference By Enumeration to calculate probability distributions.

Exercise 2 - Probabilistic Reasoning over Time

In Exercise 2, we implement filtering, prediction, smoothing and "most likely sequence" (Viterbi) calculations for a Hidden Markov Model.

Exercise 3 - Making Decisions

Exercise 3 has no code, and is about making a decision network, drawing it in Genie, and use reasonable probability distributions and dependencies.

Exercise 4 - Decision Trees

In Exercise 4, we implement Decision Tree Learning and try to predict the survival of a given passenger on Titanic.

Exercise 5 - Artificial Neural Networks

In Exercise 5, we implement a neural network that can be used as a Perceptron, or with one hidden layer. The Backpropagation algorithm is used to calculate the gradient descent and tune the weights.