Information extraction (IE) is the process of automatically extracting structured pieces of information from unstructured or semi-structured text documents. Classical problems in information extraction include named-entity recognition (identifying mentions of persons, places, organizations, etc.) and relationship extraction (identifying mentions of relationships between such named entities). We...

We present the theory and implementation for a new automatic adaptive h re nement and dere nement method for two and three dimensional elliptic problems An exact lower error bound for dere nement is obtained theoretically in terms of the nite element solution complementing the various known upper error bounds for re nement These error bounds are used to determine where to insert and or remove m...

A quasi-conformal mapping of the parametric domain onto the underlying physical domain is used to generate a 2D structured mesh with required properties: grid line orthogonality and prescribed mesh point clustering near the domain boundary. The functions implementing the mapping are sought by solving the Dirichlet problem for the elliptic partial differential equations. An additional control fo...

We present an algorithm that accelerates the extraction of iso-surfaces from unstructured grids by avoiding the traversal of the entire set of cells in the volume. The algorithm consists of a sweep algorithm and a data decomposition scheme. The sweep algorithm in-crementally locates intersected elements, and the data decomposition scheme restricts the algorithm's worst-case performance. For dat...

D2.1 presents the STAR on 3D object digitisation, shape matching and retrieval methods, along with local shape descriptors for unstructured 3D data. The first part of this report focuses on current CH 3D object digitisation techniques and practices, including the handling of special CH object forms, technologies, typical acquisition problems related to the task at hand and an evaluation of rele...

This paper presents a parallel finite element toolbox for computing large electromagnetic devices on unstructured tetrahedral meshes, FEMAG—Fem for ElectroMagnetics on Adaptive Grids. The finite element toolbox deals with unstructured tetrahedral meshes and can solve electromagnetic eddy current problems in both frequency domain and time domain. It adopts high-order edge element methods and ref...

In this paper we will introduce a measure of saturation for unstructured texts of unknown domains. Therefore we will present the Textual Coverage Rate (TCR), a method to determine the IE coverage of unstructured texts using a given vocabulary. We advance efficiency while building vocabulary repositories tailored for given problems and ensure a certain quality of representation. Our approach, wh...

Many problems involve the prediction of multiple, possibly dependent labels. The structured output prediction framework builds predictors that take these dependencies into account and use them to improve accuracy. In many such tasks, performance is evaluated by the Area Under the ROC Curve (AUC). While a framework for optimizing the AUC loss for unstructured models exists, it does not naturally...

Problems after severe soft tissue injuries, skin grafting, and flap procedures are uncontrolled hypertrophic scarring, unstable scars, functional deficits, and aesthetic disfigurements. Ongoing swelling and lymphatic stasis are also a common problem, and do contribute to functional problems. After deep skin / soft tissue injuries, an unstructured replacement tissue is formed (scar), and does no...

We consider a number of practical issues associated with the parallel distributed memory solution of elliptic partial diierential equations using unstructured meshes in two dimensions. The rst part of the paper describes a parallel mesh generation algorithm which is designed both for eeciency and to produce a well-partitioned, distributed mesh, suitable for the eecient parallel solution of an e...