Organisation

Time Monday, 8:15h-11:30h
Room 120
E-Learning Course in Ilias
Credits 4 SWS / 6 ECTS
Exam written, 60min

Announcements

  • First lesson in SS 23: 20.03.2023

Artificial Intelligence

Maybe you know Artificial Intelligence (AI) from Sci-Fi movies? With intelligent machines that are far superior to humans?

If you expect something like this, you might be on the wrong path. First because AI is not only fictitious and future stuff. It is already pervasive in our everyday life. AI is at the heart of weather forecasts, digital assistents (Amazon Echo, Siri, Google Now), recommendation systems, web-search, fraud detection, face recognition, medical- and technical diagnosis, computer games and much more. Even though we have lots of intelligent machines and applications and AI is supposed to bring major shifts in society, you should not be afraid: AI will not reach the full breadth of human intelligence soon. Intuition, emotions, common sense and social skills are only a few factors of intelligence, which make humans superior.

The message is: AI is something practical. It builds rational agents and it provides a toolbox, which enables new and improves existing applications. This lecture provides an insight in this AI toolbox. The elements of this toolbox are algorithms and datastructures. The mathematical and technical features of these algorithms and their application fields are in the focus of this lecture. On an abstract field AI, as covered in this lecture, can be partitioned into the 4 areas depicted in the figure below. For each area the algorithms and concepts, studied in this lecture, are listed.

Jupyterbook and Gitlab Repo of this lecture

Former Documents

Gitlab Repo containing all resources of this lecture. Pull repo before each lecture for staying up to date.

Lecture Contents Additional Material
1. Introduction What is AI? Applications; Categorization; Definitions;
1a. Rational Agents Specificiation of Environment, Types of Rational Agents
2. Search and Planning Heuristic and Non-Heuristic Search and Planning; Planning in Multiplayer Games; Monte Carlo Tree Search; Genetic Algorithm Python Implementation of Search Algorithms; Solving 8-puzzle (.ipynb); (.html)
3. Knowledge Representation and Inference Knowledge, Inference, Knowledge-Graphs
4. Probabilistic Inference Probability Theory, Probabilistic Inference, Bayes Nets Bayes Net Implementation (pyAgrum) (.ipynb)
5. Machine Learning Intro Definition, Application, Categorization, Evaluation Intro Supervised Learning - Regression, Intro Supervised Learning - Classification,Video Machine Learning Intro
6. Decision Trees / Ensembles Characteristics, Entropie, Information Gain, Learning, Boosting and Bagging, Random Forests Decision Tree Learning, Ensemble Learning
7. Neural Networks 1 General concepts, SLPs and MLPs for classification and regression, Stochastic Gradient Descent, Backpropagation Learning, SLP Notebook, MLP Notebook,SLP Ex. Bin.Classif.,SLP Ex. K-ary Classif.,SLP Ex. Regression,Video SLP, MLP
8. Deep Neural Networks Deep Neural Networks in general, Convolutional Neural Networks, Layer-Types, Architectures CNN Notebook,Video Deep Learning,Convolution Filtering video
9. Unsupervised Learning Preprocessing, Similarity Measures, k-means, Hierarchical Clustering, DBSCAN
10. Reinforcement Learning Bellman Equation, Model-based Learning, Q-Learning

Exercises and further materials

Exercise Contents
Exercise 1 Search, Plan, Optimize
Exercise 2 Probability, Uncertainty, Bayes Net
Exercise 3 Entropy, Information Gain, Decision Trees
Exercise 4 Single Layer Perceptron (SLP)

Literature:

  • Artificial Intelligence: A Modern Approach (4th Edition) (May, 1st, 2021) by Stuart Russell, Peter Norvig
  • Grundkurs Künstliche Intelligenz (September, 16th, 2016) by Wolfgang Ertel
  • Introduction to Machine Learning (Adaptive Computation and Machine Learning) (01 October 2004) by Ethem Alpaydin
  • Programming collective intelligence : building smart web 2.0 applications (23 August 2007) by Toby Segaran