BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260610T085653EDT-81552wjRKU@132.216.98.100
DTSTAMP:20260610T125653Z
DESCRIPTION:NCRN Seminar\n\nSpeaker: Marc G. Bellemare\, Google Brain in Mo
 ntreal\n\nZoom: Registration Link\n\nAfter registering\, you will receive 
 a confirmation email containing information about joining the meeting.\n	\n
 	ABSTRACT:\n	Efficiently navigating a superpressure balloon in the stratosph
 ere requires the integration of a multitude of cues\, such as wind speed a
 nd solar elevation\, and the process is complicated by forecast errors and
  sparse wind measurements. Coupled with the need to make decisions in real
  time\, these factors rule out the use of conventional control techniques.
  This talk describes the use of reinforcement learning to create a high-pe
 rforming flight controller for Loon superpressure balloons. Our algorithm 
 uses data augmentation and a self-correcting design to overcome the key te
 chnical challenge of reinforcement learning from imperfect data\, which ha
 s proved to be a major obstacle to its application to physical systems. We
  deployed our controller to station Loon balloons at multiple locations ac
 ross the globe\, including a 39-day controlled experiment over the Pacific
  Ocean. Analyses show that the controller outperforms Loon's previous algo
 rithm and is robust to the natural diversity in stratospheric winds. These
  results demonstrate that reinforcement learning is an effective solution 
 to real-world autonomous control problems in which neither conventional me
 thods nor human intervention suffice\, offering clues about what may be ne
 eded to create artificially intelligent agents that continuously interact 
 with real\, dynamic environments.\n	\n	BIOGRAPHY:\n	Marc G. Bellemare leads t
 he reinforcement learning efforts at Google Brain in Montreal and holds a 
 Canada CIFAR AI Chair at Mila. He received his Ph.D. from the University o
 f Alberta\, where he developed the highly-successful Arcade Learning Envir
 onment benchmark. From 2013 to 2017 he worked at DeepMind in London\, UK\,
  where he made major contributions to deep reinforcement learning\, in par
 ticular pioneering the distributional method. Marc G. Bellemare is also a 
 CIFAR Learning in Machines & Brains Fellow and an adjunct professor at McG
 ill University.\n
DTSTART:20210318T180000Z
DTEND:20210318T190000Z
LOCATION:CA\, ZOOM
SUMMARY:Autonomous navigation of stratospheric balloons using reinforcement
  learning
URL:https://www.mcgill.ca/cim/channels/event/autonomous-navigation-stratosp
 heric-balloons-using-reinforcement-learning-329332
END:VEVENT
END:VCALENDAR