Aktuelle Themen der Intelligenten Systeme und Dienste
for questions please contact
fabian.hildebrandt@tu-dresden.de
Consultation with Prof. Brendel : on request
(please contact isd@tu-dresden.de for an appointment)
Course description
Current issues in Information systems theory and practice are dealt with. Students are able to present, discuss, analyse and evaluate these issues using scientific methods. For a practical approach, the course deals with conversational agents. Hence, this course focuses on the design, development, deployment, and evaluation of conversational agents (e.g., chatbots or voice assistants) for a given problem domain (e.g., customer service, team collaboration).
Course objectives
The aim of the course is to introduce conceptual and technical foundations of conversational agents, relevant theories of human-computer interaction, and design guidelines for different classes of conversational agents. In addition, the course will introduce the user-centered design approach adapted to the design of conversational agents, including several qualitative and quantitative evaluation approaches.
The entire course is held virtually with no physical meetings, providing a first experience for future workplace scenarios. The course is a joint offering in cooperation with University Hamburg and Saarland University (UdS, Jun.-Prof. Morana). Students will work collaboratively in virtual teams with students from the other universities (i.e., one student per university in one team).
Each semester, a new challenge is presented and students must first understand the respective problem domain in more detail (e.g., through studying literature and interviews). Next, the student teams derive user needs and requirements and create a first prototype of the conversational agent. This prototype is iteratively refined and subsequently implemented as a functional prototype using contemporary technology tools (e.g., Google Dialog Flow, Microsoft Bot Framework, Rasa, etc.). Finally, the student teams evaluate their conversational agent applying qualitative and quantitative methods (e.g., online experiments, surveys, interviews, focus groups, etc.).
The student teams document the results of each activity in a project report. The project report as well as the conversational agent prototype are the basis for the grading of the course.
Learning Objectives
After the successful participation in the course students can:
- explain conceptual and technical foundations of conversational agents.
- perform the user-centered design approach to design, develop, and evaluate a conversational agent.
- deploy a conversational agent prototype based on state-of-the-art tools and frameworks.
- evaluate the conversational agent prototype using qualitative and quantitative evaluation methods.
Course facts
- Format: voluntary E-Lecture & Project
- Language: English
- Semester: every semester
- Scope: 5 CP
Important information
The number of participants is limited to 15. First come first serve. There will be a waiting list.
A kickoff meeting with all the information for the course will be held via Microsoft-Teams. A link will be sent to all enrolled students via mail.
Lectures, exercises, general communication and all other content are provided by MS-Teams. To collaborate in your group you should have basic understanding for MS-Teams. We recommand this video: Introduction for MS-Teams
Kick-Off (26.10.2023 - 14:15): Teams Link tbd
Timetable
The seminar consists of mandatory group meetings for presentations and additional individual group meetings with the respective supervisor. A list of all mandatory dates is below.
| # | date | start | end | appointment type |
| 1 | 26.10.2023 | 14:15 | 15:45 | Kick-off |
| 2 | 07.12.2023 | 14:15 | 15:45 | All groups meeting: Interim Presentation |
| 4 | 01.02.2024 | 14:15 | 17:30 | All groups meeting: Final Presentation |
| 5 | 29.02.2024 | - | - | Submission Project Report Paper |
Course requirements
- Students should have first experiences in programming and the motivation to enhance them
- Experience in human-computer interaction and conversational agents is optional (all relevant content will be provided)