BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:CompSci Colloquium: Katherine Skinner on "Deep Learning for Ro
 bot Vision in Natural Environments"
X-WR-TIMEZONE:Mountain Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260513T093838Z
UID:tag:localist.com\,2008:EventInstance_4373295
DTSTART:20190221T210000Z
DTEND:20190221T220000Z
DESCRIPTION:Learning from the Field: Deep Learning for Robot Vision in Natu
 ral Environments\n\nABSTRACT: Field robotics refers to the deployment of r
 obots and autonomous systems in unstructured or dynamic environments acros
 s air\, land\, sea\, and space. Robust sensing and perception can enable t
 hese systems to perform tasks such as long-term environmental monitoring\,
  mapping of unexplored terrain\, and safe operation in remote or hazardous
  environments. In recent years\, deep learning has led to impressive advan
 ces in robotic perception. However\, state-of-the-art methods still rely o
 n gathering large datasets with hand-annotated labels for network training
 .\nFor many applications across field robotics\, dynamic environmental con
 ditions or operational challenges hinder efforts to collect and manually l
 abel large training sets that are representative of all possible environme
 ntal conditions a robot might encounter. This limits the performance and g
 eneralizability of existing learning-based approaches for robot vision in 
 field applications.\n\nIn this talk\, I will discuss my work to develop ap
 proaches for unsupervised learning to advance perceptual capabilities of r
 obots in underwater environments. The underwater domain presents unique en
 vironmental conditions to robotic systems that exacerbate the challenges i
 n perception for field robotics. To address these challenges\, I leverage 
 physics-based models and cross-disciplinary knowledge about the physical e
 nvironment and the data collection process to provide constraints that rel
 ax the need for ground truth labels. This leads to a hybrid model-based\, 
 data-driven solution. I will also present work that relates this framework
  to challenges for autonomous vehicles on land.\n\nBIO: Katherine Skinner 
 is a Ph.D. Candidate in the Robotics Institute at the University of Michig
 an and a member of the Deep Robot Optical Perception Lab. Her research int
 erests span robotics\, machine learning\, and computer vision\, with a foc
 us on enabling autonomy in dynamic\, unstructured\, or remote environments
 . She received a B.S.E. in Mechanical and Aerospace Engineering with a Cer
 tificate in Applications of Computing from Princeton University in 2014 an
 d an M.S. in Robotics from the University of Michigan in 2016. She has hel
 d appointments at Woods Hole Oceanographic Institution and the Australian 
 Centre for Field Robotics. She is a recipient of the NSF EAPSI Fellowship.
GEO:40.0068;-105.261499
LOCATION:Discovery Learning Center\, DLC 170
SUMMARY:CompSci Colloquium: Katherine Skinner on "Deep Learning for Robot V
 ision in Natural Environments"
URL;VALUE=URI:https://calendar.colorado.edu/event/compsci_colloquium_kather
 ine_skinner_on_deep_learning_for_robot_vision_in_natural_environments
CATEGORIES:Colloquium/Seminar
END:VEVENT
END:VCALENDAR