Pierre Haritz
Forschungsthemen
- Safe Reinforcement Learning
- Multi-Agent Learning
- Transfer Learning
- Agent Verification
Kurse
- WiSe 21/22: EINI Übung/Praktikum
- SoSe 22: Proseminar Reinforcement Learning and Bandit Algorithms, Übungen zu "Data Privacy"
- WiSe 22/23: Projektgruppe PG652 - BlueBerry4School
- SoSe 23: Projektgruppe PG652 - BlueBerry4School
- WiSe 23/24: Übungen zu "Grundlagen der Datenwissenschaft", Praktikum zu "Grundlagen der Datenwissenschaft"
- SoSe 24: Projektgruppe PG662 - Facts4Chat
Betreute Abschlussarbeiten
- Transfer von Reinforcement-Learning-Modellen auf Constrained Markov Decision Processes (Dienst, Lukas)
- Standortoptimierung durch Reinforcement Learning am Beispiel von Taxidepots (Jordan, Christian)
- Effiziente Sampling-Strategien mit Stabilitätskontrolle für komplexe Fräsprozesse (Nagendram, Mayavan)
- Generalisierende Reinforcement-Learning-Agenten für sicheres autonomes Fahren in urbanen Szenarien (Wanke, David)
- Learning Robust Power Grid Control under N-1 Safety Constraints (Bhiryani, Avisha)
Bei Interesse an Abschlussarbeiten (Bachelor, Master) nehmen Sie gerne Kontakt per Email auf.
Engagement
- Fakultätskommision für Haushalt und Struktur (HaSt)
- Reviewer: European Conference of Artificial Intelligence (ECAI '23), IEEE Intelligent Vehicles Symposium (IV '24)
Publikationen
2024
- Haritz, Pierre and Wanke, David and Liebig, Thomas. Enhancing Safety for Autonomous Agents in Partly Concealed Urban Traffic Environments Through Representation-Based Shielding. In 2024 IEEE Intelligent Vehicles Symposium (IV), Pages 1758-1763.
- Sachweh, Timon(*) and Haritz, Pierre(*) and Liebig, Thomas. Using Petri Nets as an Integrated Constraint Mechanism for Reinforcement Learning Tasks. In 2024 IEEE Intelligent Vehicles Symposium (IV), Pages 1686-1692. (*: equal contribution)
- Krone, Jan Hendrik(*) and Haritz, Pierre(*) and Liebig, Thomas. ICE-T: A Multi-Level Concept for Teaching Machine Learning. ISSEP 2024, accepted. (*: equal contribution)
2023
- Haritz, Pierre and Liebig, Thomas. Free-Energy Advantage Functions for Policy Transfer to Noisy Environments with Safety Constraints. LWDA 2023, Pages 414-423.
2021
- Haritz, Pierre and Pfahler, Lukas and Liebig, Thomas and Kotthaus, Helena. Self-Supervised Source Code Annotation from Related Research Papers. In Proceedings of the PhD Forum of the 21st IEEE International Conference on Data Mining, Pages 1083-1084, 2021. PhD Forum Best Paper Award.