Semi-supervised Learning in Network-Structured Data via Total Variation Minimization.

We provide an analysis and interpretation of total variation (TV) minimization for semi-supervised learning from partially-labeled network-structured data. Our approach exploits an intrinsic duality between TV minimization and network flow problems. In particular, we use Fenchel duality to establish a precise equivalence of TV minimization and a minimum cost flow problem. This provides a link betw...