BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:PDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0700 TZOFFSETTO:-0800 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:PST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260508T024151Z UID:7CDBDB5B-A0A1-4161-BF70-9DF859AC6E4F DTSTART;TZID=America/Los_Angeles:20260529T153000 DTEND;TZID=America/Los_Angeles:20260529T170000 DESCRIPTION:How can we transform artificial intelligence (AI) and machine l earning capabilities into reliable\, autonomous robotic systems? How can w e engineer AI systems within budget constraints\, certify them with respec t to stakeholder requirements\, and ensure that they meet the needs of the end user? Answering these questions necessitates new engineering methodol ogies for AI systems\, as well as AI algorithms that leverage the unique c haracteristics of engineering problems. In this talk\, I will begin by pre senting methods that integrate foundation models such as large language mo dels and vision-language-action models with frameworks and algorithms for verifiable sequential decision-making. I will then present compositional a pproaches to reinforcement learning\, which enable independent development and testing of separate learning-enabled modules and facilitate the relia ble deployment of their compositions in practice. Finally\, I will present control-oriented learning algorithms that combine data with prior physics knowledge\, yielding learning-enabled systems that effectively control ha rdware after mere minutes of data collection and training. Experiments on robotic hardware\, ranging from manipulators to ground vehicles to hexacop ters\, demonstrate the important role that these algorithms play in the fa st and reliable transfer of learning-driven algorithms to their target\, r eal-world operating environments.\n\nCo-sponsored by: Ryozo Nagamune | nag amune@mech.ubc.ca | Dejan Kihas | kihas@ieee.org\n\nSpeaker(s): Cyrus Near y \n\nAgenda: \nEvent Start: 3:30pm\n\nTalk and Q&A: 3:40pm\n\nEvent End: 5:00pm\n\nBldg: MacLeod Building \, Room MCLD 3038 \, 2356 Main Mall\, Va ncouver \, British Columbia\, Canada\, V6T 1Z4\, Virtual: https://events.v tools.ieee.org/m/559041 LOCATION:Bldg: MacLeod Building \, Room MCLD 3038 \, 2356 Main Mall\, Vanc ouver \, British Columbia\, Canada\, V6T 1Z4\, Virtual: https://events.vto ols.ieee.org/m/559041 ORGANIZER:kihas@ieee.org SEQUENCE:85 SUMMARY:Engineering AI Systems and AI for Engineering: Language\, Compositi onality\, and Physics in Learning-Driven Robot Autonomy URL;VALUE=URI:https://events.vtools.ieee.org/m/559041 X-ALT-DESC:Description: <br /><p class="MsoNormal">How can we transform art ificial intelligence (AI) and machine learning capabilities into reliable\ , autonomous robotic systems? How can we engineer AI systems within budget constraints\, certify them with respect to stakeholder requirements\, and ensure that they meet the needs of the end user? Answering these question s necessitates new engineering methodologies for AI systems\, as well as A I algorithms that leverage the unique characteristics of engineering probl ems. In this talk\, I will begin by presenting methods that integrate foun dation models such as large language models and vision-language-action mod els with frameworks and algorithms for verifiable sequential decision-maki ng. I will then present compositional approaches to reinforcement learning \, which enable independent development and testing of separate learning-e nabled modules and facilitate the reliable deployment of their composition s in practice. Finally\, I will present control-oriented learning algorith ms that combine data with prior physics knowledge\, yielding learning-enab led systems that effectively control hardware after mere minutes of data c ollection and training. Experiments on robotic hardware\, ranging from man ipulators to ground vehicles to hexacopters\, demonstrate the important ro le that these algorithms play in the fast and reliable transfer of learnin g-driven algorithms to their target\, real-world operating environments.</ p>\n<p class="MsoNormal">&nbsp\;</p><br /><br />Agenda: <br /><p>Event Sta rt: 3:30pm&nbsp\;</p>\n<p>Talk and Q&amp\;A: 3:40pm</p>\n<p>Event End: 5:0 0pm&nbsp\;</p>\n<p>&nbsp\;</p> END:VEVENT END:VCALENDAR