Abstract

We present here the outline of an ongoing research effort to recognize, represent, and interpret attributive constructions such as reported speech in newspaper articles. The role of reported speech is attribution: the statement does not assert some information as `true' but attributes it to some source. The description of the source and the choice of the reporting verb can express the reporter's level of confidence in the attributed material.

A new technical report on Text Mining (in German) is now available. This is a collection of reports written by students within a Hauptseminar, which was given by yours truly and Jutta Mülle at Universität Karlsruhe, Germany.

Abstract

Protein structure visualization tools render images that allow the user to explore structural features of a protein. Context specific information relating to a particular protein or protein family is not easily integrated and must be uploaded from databases or provided through manual curation of input files. We describe a mixed natural language processing and protein sequence analysis approach for the retrieval of mutation specific annotations from full text articles for rendering with protein structures.

Introduction

Biological researchers today have access to vast amounts of exponentially growing research data in a structured form within several publicly accessible databases. A large proportion of salient information is however still hidden within individual research papers, since costly manual database curation efforts are overwhelmed by the scale of new information being generated. In the domain of protein engineering, critical units of information required from the literature include: the identity of the mutated protein, the identity and position of wild type residues that are mutated, the identity of the resulting mutant residues and the impacts of the mutations on functional properties of the proteins.
Mutation Miner is a system designed to automate the extraction of mutations and textual annotations describing the impacts of mutations on protein properties (mutation annotations) from full text scientific literature. Furthermore, the system retrieves and carries out bioinformatic analyses on mutated sequences providing the mapped coordinates of mutants on a selected structure. Integration of multiple formatted mutation annotations with associated residue coordinates facilitates their rendering with structure visualization tools. We describe the architecture and tools that support Mutation Miner (Text mining-NLP, Sequence Analysis, Structure Visualization) and present performance evaluations that demonstrate the feasibility of this approach.

Introduction

Biological researchers today have access to vast amounts of exponentially growing research data in a structured form within several publicly accessible databases. A large proportion of salient information is however still hidden within individual research papers, since costly manual database curation efforts are overwhelmed by the scale of new information being generated. In the domain of protein engineering, critical units of information required from the literature include: the identity of the mutated protein, the identity and position of wild type residues that are mutated, the identity of the resulting mutant residues and the impacts of the mutations on functional properties of the proteins.
Mutation Miner is a system designed to automate the extraction of mutations and textual annotations describing the impacts of mutations on protein properties (mutation annotations) from full text scientific literature. Furthermore, the system retrieves and carries out bioinformatic analyses on mutated sequences providing the mapped coordinates of mutants on a selected structure. Integration of multiple formatted mutation annotations with associated residue coordinates facilitates their rendering with structure visualization tools. We describe the architecture and tools that support Mutation Miner (Text mining-NLP, Sequence Analysis, Structure Visualization) and present performance evaluations that demonstrate the feasibility of this approach.

Abstract

Comprehension is an essential part of software evolution. Only software that is well understood can evolve in a controlled manner. In this paper, we present a formal process model to support the comprehension of software systems by using Ontology and Description Logic. This formal representation supports the use of reasoning services across different knowl- edge resources and therefore, enables us to provide users with guidance during the comprehension process that is context sensitive to their particular comprehension task. As part of the process model, we also adopt a new interactive story metaphor, to represent the interactions between users and the comprehension process.

Keywords: Software evolution, program comprehension, process modeling, story metaphor, ontological reasoning

Abstract

Reverse Engineering is a process fraught with imperfections. The importance of dealing with non-precise, possibly inconsistent data explicitly when interacting with the reverse engineer has been pointed out before.

In this paper, we go one step further: we argue that the complete reverse engineering process must be augmented with a formal representation model capable of modeling imperfections. This includes automatic as well as human-centered tools.

We show how this can be achieved by merging a fuzzy set-theory based knowledge representation model with a reverse engineering repository. Our approach is not only capable of modeling a wide range of different kinds of imperfections (uncertain as well as vague information), but also admits robust processing models by defining explicit degrees of certainty and their modification through fuzzy belief revision operators.

The repository-centered approch is proposed as the foundation for a new generation of reverse engineering tools. We show how various RE tasks can benefit from our approach and state first design ideas for fuzzy reverse engineering tools.