Abstract We consider the general problem of learning from labeled and unlabeled data, which is often called semi-supervised learning or transductive inference. A principled approach to semi-supervised learning is to design a classifying function which is sufficiently smooth with respect to the intrinsic structure collectively revealed by known labeled and unlabeled points. We present a simple algorithm to obtain such a smooth solution. Our method yields encouraging experimental results on a number of classification problems and demonstrates effective use of unlabeled data.

preprocessing improves GSAT. In particular, we will focus on the effect of enforcing local consistency on the performance of GSAT. We will show that enforcing local consistency on uniform random problems has very little e ect on the performance of GSAT. However, when the problem has hierarchical structure

We extend the framework of Constraint Satisfaction Problems to make it more suitable for/applicable to modern constraint programming languages where both constraint satisfaction and constraint solving have a role. Some rough principles for local consistency conditions in the extended framework

of the schema to handle the more general non-binary constraints, i.e., those given as a predicate or as a set of allowed/forbidden combinations of values. In this paper, we extend that work by proposing an efficient way to compute local consistency on conjunctions of constraints (called conjunctive consistency

Abstract Local consistency techniques such as k-consistency are a key component of specialised solvers for constraint satisfaction problems. In this paper we show that the power of using k-consistency techniques on a constraint satisfaction problem is precisely captured by using a particular

Abstract. We establish optimal bounds on the number of nested prop-agation steps in k-consistency tests. It is known that local consistency algorithms such as arc-, path- and k-consistency are not efficiently par-allelizable. Their inherent sequential nature is caused by long chains of nested

We consider algorithms of the Waltz type for computing local consistency (also called arcconsistency) for constraints over numeric domains.

We introduce some new mappings of constraint satisfaction problems into propositional satisability. These encodings generalize most of the existing encodings. Unit propagation on those encodings is the same as establishing relational k-arc consistency on the original problem.

A robot exploring an unknown environmentmay need to build a world model from sensor measurements. In order to integrate all the frames of sensor data, it is essential to align the data properly. An incremental approach has been typically used in the past, in which each local frame of data

We investigate the parameterized complexity of deciding whether a constraint network is k-consistent. We show that, parameterized by k, the problem is complete for the complexity class co-W[2]. As secondary parameters we consider the maximum domain size d and the maximum number ` of constraints