One of the major problem in face verification is to deal with a few number of images per person to train the system. A solution to that problem is to generate virtual samples from...
We investigate the problem of non-covariant behavior of policy gradient reinforcement learning algorithms. The policy gradient approach is amenable to analysis by information geom...
Hausdorff metrics are used in geometric settings for measuring the distance between sets of points. They have been used extensively in areas such as computer vision, pattern recog...
Ever-increasing complexity of large-scale applications and continuous increases in sizes of the data they process make the problem of maximizing performance of such applications a...
Mahmut T. Kandemir, Seung Woo Son, Mustafa Karak&o...
We show for the first time that incorporating the predictions of a word sense disambiguation system within a typical phrase-based statistical machine translation (SMT) model cons...