WS20/21 Artificial Intelligence course at University of Kaiserslautern (Germany).
Create a dataframe with data from WG-Gesucht, a portal for shared apartments and accomodation.
- Implement search algorithms by hand (A* and MinMax with Alpha-Beta pruning).
- Implement a class node with a function that performs an alpha-beta search.
We used the data collected on Assignment 1 for training 2 machine learning models to predict price:
- Linear regression, comparing results using the normal equation vs. gradient descent.
- Fully-connected neural network using the torch library.
We use supervised learning to train a classifier to recognize type of activity from an on-body sensor (accelerometer with 3 axes).
We used multiple algorithms built-in the scikit-learn library and then we created a confusion matrix to visualize the performance of the classifier.
We train a classifier for the STL-10 dataset, using pytorch. We applied transfer learning, training on top of ResNet34, a convolutional neural network.