This paper considers approaches which rerank the output of an existing probabilistic parser. The base parser produces a set of candidate parses for each input sentence, with associated probabilities that define an initial ranking of these parses. A second model then attempts to improve upon this initial ranking, using additional features of the tree as evidence. We describe and compare two appr...

Multiply sectioned Bayesian networks (MSBNs) were originally proposed as a modular representation of uncertain knowledge by sectioning a large Bayesian network (BN) into smaller units. More recently, hierarchical Markov networks (HMNs) were developed in part as an hierarchical representation of the flat BN. In this paper, we compare the MSBN and HMN representations. The MSBN representation does...

Markov logic networks (MLNs) use firstorder formulas to define features of Markov networks. Current MLN structure learners can only learn short clauses (4-5 literals) due to extreme computational costs, and thus are unable to represent complex regularities in data. To address this problem, we present LSM, the first MLN structure learner capable of efficiently and accurately learning long clause...

In real world, a scene is composed by many characteristics. Intrinsic images represent these characteristics by two components, reflectance (the albedo of each point) and shading (the illumination of each point). Because reflectance images are invariant under different illumination conditions, they are more appropriate for some vision applications, such as recognition, detection. We develop the...

Local Abstract (3rd part) Thís part contains several examples that illustrate the implementation oí the facial deductive mechanism and show how to transform information about conditional independence structure given in the form of dependency models, Bayesian or Markov networks into imsets. Another example indicates that the ťacial deductive mechanism is indeed more powerful than the semigraphoi...

Go is an ancient oriental game whose complexity has defeated attempts to automate it. We suggest using probability in a Bayesian sense to model the uncertainty arising from the vast complexity of the game tree. We present a simple conditional Markov random field model for predicting the pointwise territory outcome of a game. The topology of the model reflects the spatial structure of the Go boa...

We address the question of how a firm’s current location in the space of technological possibilities constrain its search for technological improvements. We formalize a quantitative notion of distance between technologies — encompassing the distinction between evolutionary changes (small distance) versus revolutionary change (large distance) — and introduce a technology landscape into an otherw...

We show that the usual score function for conditional Markov networks can be written as the expectation over the scores of their spanning trees. We also show that a small random sample of these output trees can attain a significant fraction of the margin obtained by the complete graph and we provide conditions under which we can perform tractable inference. The experimental results confirm that...

We explore the conditional probabilistic independences of systems of random variables (I; JjK), to read \I is conditionally independent of J given K", which are ascending (I; JjK)) (I; JjK L) or descending (I; JjK L)) (I; JjK). The resulting abstract independence structures can be equivalently described by weak families of connected sets. Using, in addition, probabilistic representations of mat...

This paper provides a complete characterization of stationary Markov random fields on a finite rectangular (nontoroidal) lattice in the basic case of a second-order neighborhood system. Equivalently, it characterizes stationary Markov fields on 2 whose restrictions to finite rectangular subsets are still Markovian (i.e., even on the boundaries). Until now, Pickard random fields formed the only ...