Elements of AI: Building AI - Advanced

The course can be completed on three difficulty levels. The following learning objectives are achieved after completing the course on the advanced level.

After the course, you are able to:

• describe different types of AI such as optimization, reasoning, and learning
• choose a suitable AI approach to solve simple tasks such as route planning, probabilistic inference, and pattern recognition
• implement a straightforward brute-force optimization algorithm
• implement simple probabilistic inference based on statistical data using the Bayes rule
• build linear regression models from data, and use the models to predict variables of interest, such as apartment prices
• use the nearest neighbor method to predict variables of interest
• use cross-validation to avoid under- and overfitting
• build and apply logistic regression and simple neural network models for prediction

• The course content consists of text, visual and interactive elements that are available online.
• The online course material is intended to be self-contained.
• Links to additional, optional readings are provided in the material.

Target group

• The course can be taken at any stage of studies.
• Suitable for non-computer science students.
• Taking the "Elements of AI: Introduction to AI" (Part I of the series) course is recommended but not required.
• The course may be offered among the university’s continuous learning courses.
  

Completion methods

• The course if completed online (fully distance learning) and there are no attendance requirements.
• The course is completed by doing exercises. Doing at least 90% of the exercises with minimum 50% correctness is required to successfully complete the course. All the exercises are automatically graded. There is no exam.

Open University and degree programme completion methods may be different.

Contents

Chapter 1: About this Course

Chapter 2: Optimization

1. Brute-force search
2. Hill climbing
3. Games

Chapter 3: Dealing with uncertainty

1. Probability fundamentals
2. The Bayes Rule
3. Naive Bayes classifier

Chapter 4: Machine learning

1. Linear regression
2. The nearest neighbor method
3. Working with text
4. Overfitting and cross validation

Chapter 5: Neural networks

1. Logistic regression
2. Neural networks
3. Deep learning

Chapter 6: Conclusions

Assessment practices and criteria

• Pass / Fail grading.
• Successful completion requires 90% completed exercises and minimum 50% correctness.
• Only one attempt is allowed in the exercises.
• If the student fails to achieve 50% correctness, they can start again.