Artifical Intelligence, ethics and engineering (Wintersemester 2021/2022)

Lecturer: Prof. Dr. Holger Giese (Systemanalyse und Modellierung) , Christian Medeiros Adriano (Systemanalyse und Modellierung) , Christian Schäffer (Systemanalyse und Modellierung)

General Information

• Weekly Hours: 4
• Credits: 6
• Graded: yes
• Enrolment Deadline: 01.10.2021 - 22.10.2021
• Teaching Form: Project seminar
• Enrolment Type: Compulsory Elective Module
• Course Language: English

Programs, Module Groups & Modules

Description

Machine Learning-based AI systems have presented impressive feats, however, the industry still faces difficult dilemmas of deploying AI components to its customers at scale. This is particularly evident for customers operating safety-critical system. From the engineering perspective, two imperatives help guide design decisions: the ethical and the explanation one. Although machines are not humans, machines are expected to obey the ethical norms that regulate human society. Ethical behavior is one of the pillars of trust, but is not sufficient, hence autonomous agents also need to explain their actions or the lack thereof.

Groups would be able to choose projects focused on certain topics.

Suggested topics (not limited):

• Autonomous Lethal Weapons
• Autonomous Driving
• Autonomous Recommender Systems (shopping, dating)
• Autonomous Identification Systems (security, medical diagnostics)
• Autonomous Support Administrative Decisions (justice, granting parole)

Ethical Judgements (sample):

• Which tasks should be restricted to humans? When not, under which conditions?
• Which biases and mistakes made by autonomous agents are morally less acceptable?
• How is blame and credit attributed:
  • when complex engineered system rely on multiple AI components
  • when humans and autonomous systems collaborate to achieve a common goal
• Are current judgments about autonomous agents atemporal or might change?

Engineering Concerns (sample):

• Which types of models can be used to effectively to express rules and uncertainty about the decisions of autonomous systems?
• How models can help to make these autonomous systems more transparent and accountable?
• How models can help to align various goals and human values?
• How to evaluate pre-trained models (e.g., GPT3) with respect to trade-offs between accuracy and fairness?

Techniques:

• Software engineering techniques from requirements engineering, analysis & modeling, verification, and validation (testing).
• Surveys, prototyping, and goal-based argumentative techniques.
• Machine learning model comparison methods.

Literature

1. IEEE, 2019, The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems, First Edition.
2. Mitchell, M., 2021,Why AI is Harder Than We Think.
3. Brundage, M., et al., 2020, Toward trustworthy AI development: mechanisms for supporting verifiable claims.
4. Morley, J., et al., 2021, Ethics as a service: a pragmatic operationalisation of AI Ethics. Minds and Machines.
5. Xiong, P., et al., 2021, Towards a Robust and Trustworthy Machine Learning System Development.
6. Cammarota, R., et al., 2020, Trustworthy AI Inference Systems: An Industry Research View.
7. Rengasamy, D., et al., 2020, Towards a More Reliable Interpretation of Machine Learning Outputs for Safety-Critical Systems using Feature Importance Fusion. 
8. Coeckelbergh, M., 2020, AI ethics. MIT Press.
9. Dennis, Louise, et al., 2016, Formal verification of ethical choices in autonomous systems, Robotics and Autonomous Systems.
10. Rudin, C., et al., 2021, Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges
11. Gohel, et al., 2021, Explainable AI: current status and future directions.
12. Karimi, et al., 2021, A survey of algorithmic recourse: contrastive explanations and consequential recommendations.
13. Motwani, S., et al., 2021, Ethics in Autonomous Vehicle Software: The Dilemmas, in Computer.
14. Bommasani, R., et al., 2021, On the Opportunities and Risks of Foundation Models.
15. Hidalgo, C. A., et al., 2021, How humans judge machines. MIT Press.

Learning

The course is a project seminar, which has an introductory phase comprising initial short lectures. After that, the students will work in groups on jointly identified experiments applying specific solutions to given problems and finally prepare a presentation and write a report about their findings concerning the experiments.

We will organize this project seminar as a hybrid event, allowing students to participate online as well as in person.

After the registration with the Studienreferat via Moodle and before the first meeting, we will query the participants to check who prefers which format. Generally, we recommend in-person meetings for the introductory meeting and the final presentations and discussion, and an online format for intermediate lectures and project meetings, but we will try to accommodate all wishes. If you have questions, please contact christian.adriano(at)hpi.de

Examination

We will grade the group's experiments (60%), reports (30%), and presentations (10%). Participation in the project seminar during meetings and other groups' presentations in the form of questions and feedback will also be required.

Dates

If you are interested in this course, please register with the Studienreferat (via Moodle poll) until 22.10. We consider your registration an expression of interest and will allow you to cancel your registration after the introductory meeting.

If you have any question on the course organization or want to register after the Studienreferat’s deadline, please contact christian.adriano(at)hpi.de

Start date: November 2 at 09:15

Room (for in-person): A 2.2

Zoom credentials (for online participants): link

Task assignments date: On November 23 and 24, we will jointly discuss the project tasks based on the topics and the students' individual interests.

Zurück