02461 Introduction to intelligent systems

Rapporteur: Mikkel Schmidt

Exam: Written examination and reports
Written exam (weight 40%) and individualized group report (weight 60%)
Participation in project work activities are mandatory.

Exam duration: 2 hours

Aid: One handwritten page (provided or own notes?)

Evaluation: 7 step scale, external examiner

General course objectives

To give the participants a basic knowledge of

  • defining aspects of intelligent systems,
  • applications of intelligent systems in image, audio, text and game data,
  • computational tools for artificial intelligence, and
  • engineering applications of intelligent systems.

Learning objectives

  • Describe key components of intelligent systems: Sensing and active data collection, machine learning, statistical evaluation and communication
  • Discuss the role of AI tools in application domains such as bio-medicine, business and commerce, information retrieval and social media
  • Discuss safety and ethical challenges in AI. Biases and stereotypes, privacy and societal impact.
  • Apply real-time AI tools to data such as image, audio, text and games. Discuss performance obtained in individual and classroom experiments
  • Use techniques for evaluation of performance and basic debugging of AI.
  • Apply scientific Python programming tools including Jupyter notebooks, Numpy, and Pytorch
  • Apply tools for managing of files and programs in the terminal
  • Apply tools for managing programming projects including version control
  • Evaluate and provide feedback for the work of other students

Content

The course provides a general introduction to AI and its tools. The course is based on a set of AI tools with applications in image, audio, text and games. A first motivating introduction to signals, machine learning, visualization and computational tools necessary for engineering AI systems. Discussion of ethics, privacy and societal impact.

Course literature

Course notes

Remarks

**** workshop notes ****

Written exam part

Change written exam to written aids

Aids allowed: One hand written page (not currently an option at DTU)

Make it a course acitvity to make the notes and do mock exams

Project work

All AI aids are allowed, and it works okay.

Checkpoints during the period - mandatory to show up x number of times for feedback.

Require hand in code also - as part of the assignment/report.

More strict report format? Word counts for each section + number of figures/tables?

Texcount on Overleaf?

Exam platform

DTU must provide some service

  • Controlled digital platform
  • Scanning service

Other ideas

Can we change the course to not be graded.

Make a report where each section is just the prompt

Make a written exam (no aids) after handing in a report, asking to summarize. Integrate prompting into teaching?

Could they compare own solution with AI solutions?