We address the problem of unsupervised domain adaptation under setting generalized target shift (joint class-conditional and label shifts). For this framework, we theoretically show that, for good generalization, it is necessary to learn a latent representation in which both marginals distributions are aligned across domains. sake, propose learning that minimizes importance weighted loss source...

We consider the problem of fairly matching the left-hand vertices of a bipartite graph to the right-hand vertices. We refer to this problem as the optimal semimatching problem; it is a relaxation of the known bipartite matching problem. We present a way to evaluate the quality of a given semi-matching and show that, under this measure, an optimal semi-matching balances the load on the right han...

Abstract Consider two independent Poisson point processes of unit intensity in the Euclidean space dimension d at least 3. We construct a perfect matching between sets that is factor (i.e., measurable function configurations commutes with translations), and property distance matched configuration points has tail distribution decays as fast possible magnitude, namely, $$b\exp (-cr^d)$$ <mml:math...

Let G = (V;E) be an undirected graph. Given an odd number k = 2l + 1, a matchingM is said to be k-optimal if it does not admit an augmenting path of length less than or equal to k. We prove jM j jM j(l+1)=(l+2), whereM is a maximum cardinality matching. If M is not already (k+ 2)-optimal, using M , in O(jEj) time we compute a (k+2)-optimal matching. We show that starting with any matching, repe...