I have attempted to revise the assessment guide in light of the extra-ordinary situation at the 2024 delivery. However, the revision is limited by the Learning Outcomes stated in the Module Description, which is absolutely authoritative.

Overview

1. The exam is viva voce.
2. Duration Normally 20 min. May extend up to a maximum of 30 min. if necessary.
3. The student manages the first seven minutes of the exam to make their case to demonstrate that the Learning Outcomes have been achieved.
4. The student is free to bring any material desired to serve as a demonstration.
   However, it is the candidate’s understanding which will be assessed, and not prepared materials, and the student must expect in-depth questions about their demonstrations.
5. The remaining 2/3 of the exam is for further questions from the examiners to make sure that all the learning outcomes are covered.

Learning Outcomes

Knowledge: Upon completion of the course, students should be able to

• K1 describe AI in terms of the analysis and design of intelligent agents or systems that interact with their environments
• K2 explain relevant AI terminology, models, and algorithms used for problem solving, as well as limitations and risks.

Skills: Upon completion of the course, students should be able to

• S1 model problems in suitable state space depending on choice of solution method
• S2 simulate models and solve problems by means of AI methods, e.g., search algorithms or computational intelligence
• S3 analyse models, AI methods, and simulation results

General competence: Upon completion of the course, students should be able to

• C1 read and understand scientific publications and textbooks on AI and reformulate the presented problems, choice of methods, and results in a short, concise manner
• C2 discuss and communicate advantages and limitations of selected AI methods for problem solving,

Interpretation of Grades and Contents

Main Topics in the Taught Programme

1. Intelligent Agents
   • What is intelligence? Rationality?
   • The role of Intelligent Agents
   • K1, K2
   • Essentially covered in January
2. Search Algorithms (mostly based on tree traversal)
   • K2, S1, S2, S3, C1
   • Essentially covered in January and February
3. Genetic Algorithms (and optimisation)
   • S2, S3
   • This was covered in greater depth than planned, including multi-objective optimisation.
4. Machine Learning and Reinforcement Learning
   • S2, S3, C2
   • Reinforcement learning was not covered, but some other areas of machine learning saw increased coverage.
   • It is advantageous to consider reinforcement learning to link link search algorithms and state machines (2) with Machine Learning, but it does not have to be emphasised as much as originally planned.
   • Not the importance of statistical testing in all applications of Machine Learning for C2.
5. Ethics and Philosophy
   • Particularly discussion of «advantages and limitations» (C2)
   • This was not covered in intended depth, but is absolutely critical for C1 and C2 and thus cannot be neglected.
   • C1, C2

The role of the project

In April the module focused on a project. Candidates are free to use this project as a basis for the seven-minute presentation, but they have to focus on aspects which high-light learning objectives. In particular,

• S1 - how do you model the problem?
• S3 - how do you analyse the problem and results?
• C2 - what can be achieved and what are the limitations?
• C2 - be prepared to discuss challenges and dilemmas on the spot

A prototype or other deliverable is not sufficient. You have to demonstrate that you understand how to solve the problem and similar problems in the future.

Regardless of whether this project is used as a basis, the exam will cover several of the five main topics, and the candidate has to be prepared to discuss all five.

Interpretation of Grade C

The Grade C represents solid working knowledge of the most important techniques in the syllabus, including the skills to select, implement, and evaluate the techniques, and an awareness of limitations and ethical caveats. This is the level required to carry out independent work in the field.

The C Candidate demonstrates

• coverage of all the learning outcomes
• cursory knowledge of each of the five main topics
• depth understanding of a couple of favourite algorithms spanning at least two of the three topics 2/3/4.
  • depth understanding requires the skills to model and analyse the problem and solution, and take active part in discussion on the spot, as discussed under the project above.
  • depth understanding entails the necessary skills to implement and test a solution.
  • depth understanding entails sufficient theoretical understanding to vary and adapt algorithms and solution techniques to different use cases
  • the candidate should also be aware of relevant applications of the algorithms studied
• the reflection to relate knowledge and competencies to own needs and future career.
  • Why is the contents of this module worth learning?

Interpretation of Grade A

The Grade A represents broad and deep understanding exceeding expectations, as it says in the national guidelines:

An excellent performance, clearly outstanding. The candidate demonstrates excellent judgement and a very high degree of independent thinking.

In addition to the requirement for a C, an A Candidate demonstrates

• depth understanding of algorithms from all the three main topics 2/3/4.
• solid theoretical understanding spanning all five main topics

Note that this level of understanding requires reading beyond what we have discussed in class. Still, students are not expected to cover all the algorithms discussed, but those algorithms that are studied should be very well understood.

Interpretation of Grade E

The Grade E represents superficial working knowledge, sufficient to make a useful contribution to a team, but insufficient for indepedent work.

The E Candidate demonstrates

• some understanding towards each of the learning outcomes
• understanding and experience with a couple of favourite algorithms,
  • representing at least two of the three topics 2/3/4.
  • with sufficient knowledge and skill to implement solutions
• knowledge of relevant applications of the algorithms studied

Grades D and B

Grades D and B are intermediate grades, showing some characteristics of both the lower and the higher grade.

Format of the Exam

The candidate has the first seven minutes to make a case for a grade, demonstrating the skills and competencies required. It is particularly important in this part of the exam to cover favourite algorithms and other highlights.

The student may bring demonstrations to use in the first part, but assessment is based on the understanding of the techniques demonstrated and not the quality of the demonstration itself.

Do not make slide shows to display theory more advanced than what you are able to discuss in conversation. In an exam you want to appear to know a lot more than what you are able to tell in the time allotted! That image is immediately broken if you have more on the slides than you have in your head.

The remaining two thirds of the exam is managed by the examiner. This is used to make sure that the student and the examiners agree that all the necessary material is covered.

Note that it is the student’s responsibility to promote the highlights of their understanding, including depth on some selected topics. It is the examiner’s responsibility that sufficient breadth is covered.