Course syllabus
Course-PM
SSY236 Decision-making for autonomous systems lp2 HT22 (7.5 hp)
The course is offered by the department of Electrical Engineering
Aim
The purpose is to introduce the students to Artificial Intelligence methods to enable high-level decision-making in robotic systems. The topics of designing a learning system and explainable AI are chosen to enhance the knowledge of students towards the new trends in robotics. This will allow students to have a better understanding of more demanding material to solve real-life problems.
Prerequisites
SSY165 Discrete event systems or similar
ESS101 Modeling and Simulation or similar
Programming skills (we will use C++)
Examiner
Karinne Ramirez-Amaro, karinne@chalmers.se
Lectures
Karinne Ramirez-Amaro, karinne@chalmers.se
Emmanuel Dean, deane@chalmers.se
Teaching Assistants
Maximilian Diehl, diehlm@chalmers.se
Student Representatives:
MPSYS Philip Johansson phijo@student.chalmers.se
MPSYS Viet Le vietlehung99@gmail.com
MPSYS Jonas Persson jonaspersson1998@gmail.com
MPMOB Aakash Rishi rishiaakash007@gmail.com
Read more about your role as a student representative, here: https://student.portal.chalmers.se/sv/chalmersstudier/minkursinformation/kursvardering/Sidor/Att-vara-studentrepresentant.aspx
Course purpose
This course covers the following topics:
* Introduction to embodied intelligence, as part of a Cognitive system.
* Tree-based learning approaches such as Decision trees.
* Explanation-based learning methods, commonly known as Explainable-AI (XAI). This will include an introduction to knowledge representation and reasoning methods.
* High-level robot programming. This will include the design and implementation of different modules such as perception, learning, and decision-making for an autonomous system. Most of the implementation will be done in C++, therefore basic knowledge is required.
Schedule
The following list provides suggested literature for this course. This literature is not mandatory. It is intended to provide additional support for the course. Participants may use it to acquire more detailed information about the topics covered in this course.
Machine Learning. T. M. Mitchell, McGraw-Hill.
Artificial Intelligence: A Modern Approach, S. Jonathan Russell, P. Norvig, Pearson.
Additional material will be provided via Canvas for the course.
Communication
Canvas will be used as our main communication means between students and teachers. Note, that students can also use the discussion forums to interact with other students.
Changes made since the last occasion
- The home assignment structure changed as follows:
- the number of assignments was reduced from four to three
- the grading strategy of assignments was changed, this time, the assignments with more complexity have more points.
- the workload and content of the assignments were adjusted to cover more topics that will help the students to finish their final projects
- this year, we implemented mandatory attendance for all the tutorials and the guest lecture. The students need to have at least 80% attendance to pass the course
- this year, to pass the laboratory, the students need to pass all three individual assignments
- this year, the students have the opportunity of resubmitting only the failed assignments before January 4, 2023
- this year, to pass the course, the students need to pass all the online quizzes and attend at least 80% of all the tutorials and guest lectures.
- New tutorials on ROS are provided and we will use a new system based on Dockers to work on the assignments.
- The exercise session has been revised and they will be better connected to the lecture topics.
Learning objectives and syllabus
After completion of the course, the student should be able to:
- Understand the fundamental concepts for designing a learning method to tackle autonomous system problems such as reasoning, learning and prediction.
- Analyze and apply advanced reasoning techniques. The emphasis will be on learning how to design and deploy reasoning approaches in robotic systems.
- Understand different reasoning methods such as deductive, inductive and probabilistic. Explain their applications and limitations applied to real problems in autonomous systems.
- Apply the learned concepts of the explainable Artificial Intelligence methods and assess their performance in complex situations.
Link to the syllabus on Studieportalen.
If the course is a joint course (Chalmers and Göteborgs Universitet) you should link to both syllabus (Chalmers and Göteborgs Universitet).
Examination form
The final grade of this course will be awarded via continuous assessment in the form of 1) individual assignments 2) Online Quizzes and 3) a final team project.
To pass this course, the students are required to pass
- three individual assignments
- all the online quizzes
- a team final project
Individual Assignments: To pass the laboratory assignments, you need to pass all three assignments.
The goal of each assignment is to provide practical experience in implementations of the learned AI & reasoning algorithms applied in a robotic system. Each assignment will have a set of tasks, which in total can accumulate different points as follows:
| Assignment | Max Points | Min Points |
| A01 | 6 | 3 |
| A02 | 10 | 5 |
| A03 | 14 | 7 |
Each assignment will be considered as a "pass" when it completes at least half of the assignment points (see above table). The assignments should be strictly delivered on time according to the course schedule. Any assignment submitted after the deadline will be considered as not delivered (it will be marked as failed). In total, you will have three assignments, and then the laboratory will be passed when the participant passes all three assignments.
For each assignment, you need to deliver original material in the form of code or a simulation file, depending on the assignment. The code must be accompanied by a short report (e.g. readMe file) that describes the delivered solution and how to run it (in the case the delivered material requires custom initialization). All submissions will be expected via canvas.
Online Quizzes. In addition, the active participation of the students in the lectures and tutorial sessions will be evaluated via online (Canvas) quizzes. The students are required to answer all the online quizzes.
Final project: An important part of the course is a final project that combines all the main topics covered in the lectures as well as the assignments and tutorials of this course. The final project will be performed in teams of 2-3 people.
Note: even when you do the project work in teams, you will be examined individually. Every project member should be involved in all parts of the project and the final presentation. Failing to actively contribute to all parts of the project might result in a failed grade.
For the final project, each member of the team can earn between 0-60 points. A passing grade for the project requires at least 30 points. Each team will present their obtained results for the final project, where each member of the team needs to present his/her contributions. Additionally, each team will deliver a final project presentation to demonstrate their obtained results.
Finally, your final grade is computed with the following formula:
Total grade= Assignment + Final project + Quizzes
See the below table to convert the number of points to the final course grade.
| Number of points | Final grade |
| 87 - 100 | 5 |
| 73 - 86.99 | 4 |
| 60 - 72.99 | 3 |
| 0 - 59.99 | fail |
The evaluation criteria for the final project include:
- The complexity of the project
- The final presentation of the project
- The code and documentation of the final implementation of the project
- Group evaluation report
Important note about assignments and final project: Each assignment must deliver original material, similarly for the final project:
• In the case of detected “copied” material, all the copies will be rejected (marked as not delivered).
• If the material is based on external sources (i.e. code from the internet!), the original source must be properly acknowledged in the comments of the delivered code, indicating the URL and the main differences.
• A brief report should be delivered with each assignment (brief solution description and mini-howTo instructions).
Individual Responsibilities
Every member of a project group should take a full part in every aspect of the project and be ready to answer any question. It is every group member's responsibility to make sure that every group member fully participates in the project. If a group has a member that does not contribute and follow the project, the group members should contact the teacher responsible for the group. Freeriders are not accepted and risk failing the project. Individual group evaluation at the end of the project. Members of the same group might get a different number of project points.
Academic work relies on respect for other people's knowledge and ideas. By representing someone else's work or ideas as your own you commit a serious offence and violate Chalmers rules. By violating the conditions stipulated in the document "Rules of Discipline" it is possible to be suspended or expelled from Chalmers. More information can be found in the documents below.
Every group has two project meetings. Attendance is mandatory. All students should be present in the final project presentation. In case you get sick you have to contact the examiner.
Course summary:
| Date | Details | Due |
|---|---|---|