Artificial Intelligence will change all processes of life substantially. Understanding the potentials and pitfalls is essential for applying or rejecting AI-based solutions in every situation of your future career. Introductions to AI are usually targeted at Statisticians or Computer Scientist. In this course I would like to lay to teach the conceptual foundations without the low level engineering or math problems associated, while still staying as actionable as possible. Wish us luck :-)
This course is target at Healthcare Professionals with basic skills in an interpreted programming language (R, Python).
Course Goals: - Understand and apply the basics of Knowledge representation - Enable Specification of Software and AI Needs, Basic Implementation Skills - Understand opportunities and limitations of ML and AI in Public Health
A. Knowledge Representation
- Terminologies, Vocabularies and Taxonomies
- Ontologies and Knowledge Graphs
- Linked Data and Knowledge Representation Languages
- Building Knowledge-based Systems
B. Logic, Inference, and Statistical Learning
- Overview on supervised and unsupervised ML methods
- Liner Regression
- Classification
- GAN, Deep Learning
- Generative AI
Lecture:
- From Data to Knowledge with Cognitive Science
- Syntax, Semantics, Pragmatics
Hands-on Activation: Visual Introduction to Machine Learning
- http://www.r2d3.us/visual-intro-to-machine-learning-part-1/
- http://www.r2d3.us/visual-intro-to-machine-learning-part-2/
Lecture:
- Introduction and History of AI: Chapter 1
Lecture:
- Intelligent Agents: Chapter 2
- Rationality and Environments
- Agents: Simple, Model-based, Goal-based, Utility-based, Learning
Exercise:
- Sudoku (1/2): https://medium.com/@co.2020.prkude/formulation-of-csp-problem-sudoku-puzzle-7d5e1d547382, https://github.com/norvig/pytudes/blob/main/ipynb/Sudoku.ipynb
Lecture Chapter 3-6:
- Search (8-Block)
- Complex Search (8-Queens)
- Constraint Satisfaction (Sudoku)
- Adversarial Search and Games (Advanced Chess, Backgammon as Stochastic Game)
Exercise:
- Sudoku (2/2) as a Constraint Satisfaction Problem, Backtracking Algorithm in Depth
Lecture Chapter 7a:
- Propositional Logic
- Semantics
- A simple Knowledge Base
Exercise:
- Wumpus World
Lecture Chapter 7b:
- Inference
- Forward and Backward Chaining: https://builtin.com/artificial-intelligence/forward-chaining-vs-backward-chaining
Exercise:
- Wumpus World
Lecture Chapter 10:
- Knowledge Representation
- History and Theory of Taxonomies, Ontologies and Semantic Networks
- Linked Data and Languages for Knowledge Representations (OWL, RDF)
Exercise:
- Building an Ontology
Recap and Interactive:
- Building an Ontology
- Ideating Projects
Lecture:
- Inference
- SPARQL Queries
Exercise:
- Querying DBPedia
-
Statistical Learning (Overview and Basic understanding)
- Supervised vs. Unsupervised Learning
- Interpretability vs. Flexibility
- Common Errors (Overfitting,...)
-
Linear Regression
-
Classification
Reading:
- Mining Ontologies from Text
- Text-Mining
- Applications in Life-Sciences
- Applications in Life Sciences
- Regulatory Restrictions on Deep Learning in Healthcare
- Applications in Life Sciences
- Regulatory Restrictions
- Discussion: Personalized Medicine and Generative AI
- A | Norvig, Russel (2021). Artificial Intelligence - A Modern Approach: https://ebookcentral.proquest.com/lib/th-deggendorf/reader.action?docID=6563527&ppg=227
- B | Hastie et al. (2021). An Introduction to Statistical Learning: https://www.statlearning.com/