Organizers

• Goran Radanovic: Instructor
• Adish Singla: Instructor
• Rati Devidze: Teaching assistant
• Stylianos Triantafyllou: Teaching assistant

Timeline and updates

• Until 26 Oct 2020: Register for the seminar course allocation at https://seminars.cs.uni-saarland.de.
• 2 Nov 2020: We have a new mailing list that includes all the four organizers/tutors. To reach out to us, you should send an email to multiagentrl-w20-tutors@mpi-sws.org (instead of contacting individuals).
• Until 30 Nov 2020: After you have been allocated a slot in the seminar, you then need to register for the seminar course examination at UdS. You should check with UdS when the examination registration starts and ends.
• 10 Nov 2020: Paper assignment for reading and writing reports will be sent to students. A total of six papers will be assigned to each student for which they will be writing reports.
• 30 Nov 2020: Reports for the first two papers are due.
• 20 Dec 2020: Reports for the next two papers are due.
• 10 Jan 2021: Reports for the last two papers are due.
• 15 Jan 2021: Paper assignment for presentations is sent to students. One paper is assigned to each student that they will be presenting.
• 15 Feb 2021: Presentation slides are due.
• Between 20 Feb to 20 Mar 2021: Final presentations will take place where each student will present their assigned paper. The exact dates will be finalized in discussion with enrolled students.

Course structure

Reading research papers

• Each student will be assigned a total of six research papers (two papers per topic). The complete list of papers is provided below, and the assignments will be done by choosing papers at random from this list. You will receive this assignment from us via email.
• For each of the assigned paper, you will have to write a two-page report.
• Each report should be submitted as a PDF file via sending an email to multiagentrl-w20-tutors@mpi-sws.org. You should name your PDF files as lastname_#.pdf (i.e., lastname_1.pdf, lastname_2.pdf, lastname_3.pdf, lastname_4.pdf, lastname_5.pdf, and lastname_6.pdf).
• Reports should be written in latex using NeurIPS style files
  • https://neurips.cc/Conferences/2020/PaperInformation/StyleFiles.
  • You should use the option \usepackage[preprint]{neurips_2020} (Non-anonymous preprints).
  • The report should include the paper title and your name at the top.
• Structure the report as an extended review, e.g.,
  • Summarize the paper.
  • Write down main strengths of the paper.
  • Write down main weaknesses of the paper.
  • Write down ways in which this paper could be improved.
  • Write down ideas in which this paper could be extended.
• These six reports will correspond to 50% of the final score.

Presentations

• Each student will be assigned a paper for presentation. This paper will be selected from the list of six papers assigned for writing reports.
• You will have to prepare a presentation of 25 mins. You will have the possibility to get feedback on your slides before the final submission.
• At the end of the semester, you will give a final presentation. We will block about 6 hours of time for the presentations. The exact dates will be finalized in discussion with enrolled students. Attendance to the final presentations will be mandatory.
• The slides and presentation will correspond to 50% of the final score.

List of research papers

Game playing and social interaction

• Multi-agent Reinforcement Learning in Sequential Social Dilemmas
  by J.Z. Leibo, V. Zambaldi, M. Lanctot, J. Marecki, and T. Graepel, at AAMAS 2017.
• Counterfactual Multi-Agent Policy Gradients
  by J.N. Foerster, G. Farquhar, T. Afouras, N. Nardelli, and S. Whiteson, at AAAI 2018.
• Simplified Action Decoder for Deep Multi-Agent Reinforcement Learning
  by H. Hu and J.N. Foerster, at ICLR 2019.
• Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning
  by N. Jaques et al., at ICML 2019.
• Learning to Teach in Cooperative Multiagent Reinforcement Learning
  by S. Omidshafiei et al., at AAAI 2019.

Human-machine collaboration

• Multi-view Decision Processes: The Helper-AI Problem
  by C. Dimitrakakis, D.C. Parkes, G. Radanovic, and P. Tylkin, at NeurIPS 2017.
• Machine Theory of Mind
  by N. C. Rabinowitz et al., at ICML 2018.
• On the Utility of Learning about Humans for Human-AI Coordination
  by M. Carroll et al., at NeurIPS 2019.
• Towards Deployment of Robust Cooperative AI Agents: An Algorithmic Framework for Learning Adaptive Policies
  by A. Ghosh, S. Tschiatschek, H. Mahdavi, and A. Singla, at AAMAS 2020.
• Learner-aware Teaching: Inverse Reinforcement Learning with Preferences and Constraints
  by S. Tschiatschek, A. Ghosh, L. Haug, R. Devidze, and A. Singla, at NeurIPS 2019.

Robustness, accountability, and safety

• Tactics of Adversarial Attack on Deep Reinforcement Learning Agents
  by Y.C. Lin et al., at IJCAI 2017.
• Inverse Reward Design
  by D. Hadfield-Menell, S. Milli, P. Abbeel, S.J. Russell, and A. Dragan, at NeurIPS 2017.
• Continuous Adaptation via Meta-Learning in Nonstationary and Competitive Environments
  by M. Al-Shedivat et al., at ICLR 2018.
• Policy Teaching via Environment Poisoning: Training-time Adversarial Attacks against Reinforcement Learning
  by A. Rakhsha, G. Radanovic, R. Devidze, X. Zhu, and A. Singla, at ICML 2020.
• Using Reward Machines for High-Level Task Specification and Decomposition in Reinforcement Learning
  by R. T. Icarte, T. Q. Klassen, R. Valenzano, and S. A. McIlraith, at ICML 2018.