BEGIN:VCALENDAR
VERSION:2.0
PRODID:OpenCms 21.0.11
BEGIN:VTIMEZONE
TZID:Europe/Berlin
X-LIC-LOCATION:Europe/Berlin
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700329T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701025T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
END:STANDARD
END:VTIMEZONE				
BEGIN:VEVENT
DTSTAMP:20260127T113901
UID:64878d1f-fb6c-11f0-b1bb-000e0c3db68b
SUMMARY:Talk of Prof. Lars Lindemann
DESCRIPTION:Abstract\nAccelerated by rapid advances in machine learning and AI, there has been tremendous success in\nthe design of learning-enabled autonomous systems in areas such as autonomous driving and robotics.\nThese exciting developments are accompanied by new fundamental challenges that arise regarding the\nsafety and reliability of these increasingly complex systems due to imperfect learning, system\nunknowns, and uncertain environments. Conformal prediction (CP) — a statistical tool for\nuncertainty quantification — has gained popularity due to its ability to deal with these\nchallenges. However, CP-based safety guarantees assume i.i.d. data, an assumption that is violated\nwhen system changes induce shifts in the underlying data distribution.\nIn this talk, I will provide new insight to design safe controllers under distribution shifts\nusing robust CP. I will begin by advocating for the use of CP due to its simplicity, generality,\nand efficiency as opposed to existing optimization-based verification techniques. I will then\nprovide an introduction to CP and summarize existing work that uses CP to design probabilistically\nsafe controllers in dynamic environments. Subsequently, we will look into interactive settings\nwhere the system’s behavior may change the environment's behavior, and vice versa. This circular\ndependency creates an interaction-driven distribution shift that invalidates existing safety\nguarantees. To deal with this chicken-and-egg problem, we propose an iterative framework that\nepisodically updates the controller while robustly maintaining safety guarantees by quantifying the\npotential impact of a controller update on the environment's behavior. We realize this via\nadversarially robust CP where we perform a regular CP step in each episode using observed data\nunder the current controller, but then transfer safety guarantees across controller updates by\nanalytically adjusting the CP result to account for distribution shifts. Lastly, we will discuss\nways to deal with more general distribution shifts that go beyond this interactive setting using\nadaptive and distributionally robust CP.  \nBiographical Information\nLars Lindemann is currently an Asst. Professor for Algorithmic Systems Theory in the Automatic\nControl Laboratory at ETH Zürich. From 2023 to 2025 he was an Asst. Professor in the Thomas Lord\nDept. of Computer Science at the University of Southern California. From 2020 to 2022 he was a\nPostdoctoral Fellow in the Dept. of Electrical and Systems Engineering at the University of\nPennsylvania. He received his Ph.D. degree in Electrical Engineering from KTH Royal Institute of\nTechnology in 2020. His research interests include systems and control theory, formal methods,\nmachine learning, and autonomous systems. Prof. Lindemann received the Outstanding Student Paper\nAward at the 58th IEEE Conference on Decision and Control and the Student Best Paper Award (as an\nadvisor) at the 60th IEEE Conference on Decision and Control. He was finalist for the Best Paper\nAward (as an advisor) at the 2024 International Conference on Cyber-Physical Systems, the Best\nPaper Award at the 2022 Conference on Hybrid Systems: Computation and Control, and the Best Student\nPaper Award at the 2018 American Control Conference.\n\n\n\n\n  
DTSTART;TZID=Europe/Berlin:20260203T160000
DTEND;TZID=Europe/Berlin:19700101T010000
LOCATION:Institute for Systems Theory and Automatic Control, , Seminar room 2.255, Pfaffenwaldring 9, 70569  Stuttgart, Campus Vaihingen 
URL;VALUE=URI:https://www.ist.uni-stuttgart.de/events/Talk-of-Prof.-Lars-Lindemann/
END:VEVENT
END:VCALENDAR