Scalable Graph-Based Learning Applied to Human Language Technology Andrei Alexandrescu Chair of the Supervisory Committee: Associate Research Professor Katrin Kirchhoff Electrical Engineering Graph-based semi-supervised learning techniques have recently attracted increasing attention as a means to utilize unlabeled data in machine learning by placing data points in a similarity graph. However, ...

Graph-based semi-supervised classification heavily depends on a well-structured graph. In this paper, we investigate a mixture graph and propose a method called semi-supervised classification based on mixture graph (SSCMG). SSCMG first constructs multiple k nearest neighborhood (kNN) graphs in different random subspaces of the samples. Then, it combines these graphs into a mixture graph and inc...

synthetic plants as well as wild crop species are valuable genetic resourses. to determine the genetic variation in triticale, tritipyrum as well as wheat lines, genomic dna of this amphiploid including thinopyrum bessarabicum and durum wheat were amplified using 32 random and 22 semi random primers. primers that could produce clear, polymorphic and repetitive bands are used for determination o...

In this note we prove an explicit formula for the lower semicontinuous envelope of some functionals defined on real polyhedral chains. More precisely, denoting by H : R → [0,∞) an even, subadditive, and lower semicontinuous function with H(0) = 0, and by ΦH the functional induced by H on polyhedral m-chains, namely

A graph is semi-symmetric if it is regular and edge transitive but not vertex transitive. The 3and 4-valent semi-symmetric graphs are wellstudied. Several papers describe infinite families of such graphs and their properties. 3-valent semi-symmetric graphs have been completely classified up to 768 vertices. The goal of this project is to extend this work to 5-valent semi-symmetric graphs. In th...

In this paper, we present a novel semi-supervised learning framework based on `1 graph. The `1 graph is motivated by that each datum can be reconstructed by the sparse linear superposition of the training data. The sparse reconstruction coefficients, used to deduce the weights of the directed `1 graph, are derived by solving an `1 optimization problem on sparse representation. Different from co...