Domain Adaptation

Principal Investigator(s): 
Trevor Darrell

ICSI researchers are investigating the fundamental problem of visual domain adaptation, or how to deal with the most common scenario “What you see is not what you get.” When test data and training data come from differing distributions (or unsupervised methods are employed with non-stationary distributes), conventional approaches to machine learning often perform very poorly. They have been exploring several approaches to this problem, including those based on conventional feature spaces that are transformed based on a learned adaptation to overcome a domain shift. They also have shown recently that recently introduced deep convolutional architectures trained on large amounts of related data are inherently invariant to many common domain shifts, yet in many practical situations, significant shifts remain. In addition to pioneering new methods to learn domain shifts, they are investigating new paradigms and problem statements, including “continuous domain adaptation,” where they consider how to model, e.g., the effect of time-varying domain non-stationary, such as weather variation in webcams.