The amount of information preserved in Portuguese archives has increased over the years. These documents represent a national heritage of high importance, as they portray the country’s history. Currently, most Portuguese archives have made their finding aids available to the public in digital format, however, these data do not have any annotation, so it is not always easy to analyze their content. In this work, Named Entity Recognition solutions were created that allow the identification and classification of several named entities from the archival finding aids. These named entities translate into crucial information about their context and, with high confidence results, they can be used for several purposes, for example, the creation of smart browsing tools by using entity linking and record linking techniques. In order to achieve high result scores, we annotated several corpora to train our own Machine Learning algorithms in this context domain. We also used different architectures, such as CNNs, LSTMs, and Maximum Entropy models. Finally, all the created datasets and ML models were made available to the public with a developed web platform, NER@DI.
The ability to extract and merge data that from documents (or databases) of different types, in order to acquire knowledge from a vast repository of information, is of unquestionable value. However that desirable integration is not an easy task. Different approaches can be followed to achieve it, ranging from the merge of resources (implying their conversion to a common format) till the fusion of the extracted parts. The idea is to interoperate those resources keeping them independent, without changes or transformations, creating over them an integration layer that gives us a general overview, as the information slices were gathered. This is possible creating a semantic network, or a conceptual map, over the resources, which relates data items among them mapping each one to its different occurrences in the repository; formally speaking, that conceptual map corresponds to the ontology that describes the knowledge we want to acquire. In this paper, we introduce Metamorphosis, a Topic Maps oriented environment to extract data from heterogeneous information repositories and to generate a browser and conceptual navigator for the extracted knowledge.
1 January 2009 in International Journal of Reasoning-based Intelligent Systems
Topic map constraint language (TMCL) provides a means to express constraints on topic maps conforming to ISO/IEC 13250. In this article, we will use a test suite and show, step-by-step, the way we handled several kinds of topic maps constraints in many different instances in order to answer questions like: Do they do the same job? Are there some kinds of topic maps constraints that are easier to specify with one of them? Do you need different background to use the tools? Is it possible to use them in similar situations (the same topic maps instances)? May we use them to produce an equal result? How do AsTMa!, OSL, Toma and XTche relate to TMCL? What kind of constraints each one of these three cannot specify? We will conclude this paper with a summary of the comparisons accomplished between those topic maps constraint languages over the use case proposed.
Everyday a huge number of new information resources are linked to the web. This way the web is growing very fast, making search tasks more and more difficult with worse results. To solve the problem several initiatives were undertaken and a new area of research and development emerged: the one called Semantic Web.When we refer to the semantic web we are thinking about a network of concepts. Each concept has a group of related resources and can be related to other concepts; we can then use this concept network to navigate among web resources or simply among information resources. From the undertaken initiatives one became an ISO standard: Topic Maps ISO 13250. The aim of this paper is to introduce a Topic Map (TM) Builder, that is a processor that extracts topics and relations from instances of a family of XML documents.A TM-Builder is strongly dependent on the resources structure. So, to extract a topic map for different collections of information resources (sets of documents with different structures) we have to implement several TM-Builders, one for each collection. This is not very easy! To overcome this inconvenient we have created an XML abstraction layer for TM-Builders that enables us to specify the topic map we want to build from a concrete family of resources, in order to generate automatically the intended extractor. To describe that process, i.e. the extraction of knowledge from XML documents to produce a TM, we present a language to specify topic maps for a class of XML documents, that we call XSTM (XML Specification for Topic Maps). We also discuss a XSL processor that automatically generates the Extractor from its formal specification written in XSTM, the XSTM-P.