nDiscovering Hidden Variables
nCan we use Active Learning to guide us?
n
nCoping with Missing or Unlabeled Data
nUse unlabeled pool to improve SVM performance
nAllow incorporation of partial experimental results
n
nDifferent action costs
nPerforming some experiments more costly than others
nExtra cost to observe some domain variables
nE.g., should we perform NMR scan on this patient or
not?
n
nScaling up to even larger domains
nImages databases with millions of images
n