Materiaalit
Muu
Ilmoittautuminen ja opintomaksu
Kuvaus
- 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.
If you are completing DEFA (Digital Education For All) studies, Elements of AI: Building AI is one of the courses that can be completed for application purposes. Please see more information in Finnish.
The programming exercises require beginner level familiarity with the Python programming language. No other prerequisites exist.
The course can be completed on three difficulty levels. The following learning objectives are achieved after completing the course on the advanced level. On the intermediate level, the same objectives are partially achieved.
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
Chapter 1: About this Course
Chapter 2: Optimization
- Brute-force search
- Hill climbing
- Games
Chapter 3: Dealing with uncertainty
- Probability fundamentals
- The Bayes Rule
- Naive Bayes classifier
Chapter 4: Machine learning
- Linear regression
- The nearest neighbor method
- Working with text
- Overfitting and cross validation
Chapter 5: Neural networks
- Logistic regression
- Neural networks
- Deep learning
Chapter 6: Conclusions
- 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.
- 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.
Taking this course does not give students access to a University of Helsinki user account. If you want to view and share information about your studies after completing the course, sign up for the Oma Opintopolku -service maintained by the Finnish National Agency for Education.
Questions regarding the course:
- You can find frequently asked questions (FAQ) in the online learning environment.
- The online learning environment: mooc@cs.helsinki.fi
- Open University course enrollment: avoinyo-tietojenkasittelytiede@helsinki.fi
- DEFA studies: DEFA-help@cs.helsinki.fi.
- Other questions about the course: Elements of AI Spectrum community
- Content of the course: teacher in charge of the course: teemu.roos@cs.helsinki.fi
Basics
The course is
- a MOOC (Massive Open Online Course) that is completed online (fully distance learning) on the online learning platform.
- There are no attendance requirements.
- You can study the contents at your own pace.
Exercises
- 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.
Difficulty levels
You can complete the course on one of the three difficulty levels. The number of credits depends on the chosen difficulty level.
- 0 credits: no credits are offered for the beginner level
- 1 credits: intermediate level (modifying code)
- 2 credits: advanced level (writing code)
Teemu Roos. Elements of AI: Building AI course is created by Reaktor and the University of Helsinki.
The course is part of the subject studies in Computer Science.