All Blogs — 4 Blogposts

AI-Lab3 — Reinforcement Learning

A one page description of the lab results

By Dano Roost, Jennifer Schürch & Yves Lütjens

The aim of this lab is to implement a Reinforcement Learning algorithm that is able to provide a good solution for the Lunar Lander environment. We implemented three different approaches, namely a Deep Q Network (DQN), a REINFORCE and an Actor-Critic (AC) algorithm.

AI-Lab2 — Supervised Learning

A one page description of the lab results

By Dano Roost, Jennifer Schürch & Yves Lütjens

The aim of this lab is to train a binary classifier that labels text with hate speech or not hate speech. The baseline already achieves an accuracy of 90% but with a very unbalanced set, meaning 11% are hate speeches and the rest are non-hate speeches.

AI-Lab1 — Unsupervised Learning

A one page description of the lab results

By Dano Roost, Jennifer Schürch & Yves Lütjens

The aim of this lab is to analyse the genre of artists based only on their song lyrics, which is done using unsupervised machine learning approaches.

Artificial Intelligence for 2048

An approach to beat 2048 with AI

By Jennifer Schürch & Yves Lütjens

The game 2048 is a hard game to beat. As there are many random variables during play, one has a hard time to implement a good algorithm for it. We started off by implemented a heuristic algorithm and set the aim that it can reach 10'000 points in average.

Solving Sudoku with CSP and creating a dragnet with Datalog

By Jennifer Schürch & Yves Lütjens

Writing a program to solve any possible Sudoku riddle can rather easily be done using the Constraint Satisfaction Problem (CSP). By adding only a few constraints one can solve even the hardest Sudoku in no time. We created two different approaches and let them run against each other.