Discriminative training of the hidden vector state model for semantic parsing

In this paper, we discuss how discriminative training can be applied to the hidden vector state (HVS) model in different task domains. The HVS model is a discrete hidden Markov model (HMM) in which each HMM state represents the state of a push-down automaton with a finite stack size. In previous applications, maximum-likelihood estimation (MLE) is used to derive the parameters of the HVS model. However, MLE makes a number of assumptions and unfortunately some of these assumptions do not hold. Discriminative training, without making such assumptions, can improve the performance of the HVS model by discriminating the correct hypothesis from the competing hypotheses. Experiments have been conducted in two domains: the travel domain for the semantic parsing task using the DARPA Communicator data and the Air Travel Information Services (ATIS) data and the bioinformatics domain for the information extraction task using the GENIA corpus. The results demonstrate modest improvements of the performance of the HVS model using discriminative training. In the travel domain, discriminative training of the HVS model gives a relative error reduction rate of 31 percent in F-measure when compared with MLE on the DARPA Communicator data and 9 percent on the ATIS data. In the bioinformatics domain, a relative error reduction rate of 4 percent in F-measure is achieved on the GENIA corpus.

Ontology-based protein-protein interactions extraction from literature using the hidden vector state model

This paper proposes a novel framework of incorporating protein-protein interactions (PPI) ontology knowledge into PPI extraction from biomedical literature in order to address the emerging challenges of deep natural language understanding. It is built upon the existing work on relation extraction using the Hidden Vector State (HVS) model. The HVS model belongs to the category of statistical learning methods. It can be trained directly from un-annotated data in a constrained way whilst at the same time being able to capture the underlying named entity relationships. However, it is difficult to incorporate background knowledge or non-local information into the HVS model. This paper proposes to represent the HVS model as a conditionally trained undirected graphical model in which non-local features derived from PPI ontology through inference would be easily incorporated. The seamless fusion of ontology inference with statistical learning produces a new paradigm to information extraction.

Probabilistic resource space model for managing resources in cyber-physical society

Classification is the most basic method for organizing resources in the physical space, cyber space, socio space and mental space. To create a unified model that can effectively manage resources in different spaces is a challenge. The Resource Space Model RSM is to manage versatile resources with a multi-dimensional classification space. It supports generalization and specialization on multi-dimensional classifications. This paper introduces the basic concepts of RSM, and proposes the Probabilistic Resource Space Model, P-RSM, to deal with uncertainty in managing various resources in different spaces of the cyber-physical society. P-RSM’s normal forms, operations and integrity constraints are developed to support effective management of the resource space. Characteristics of the P-RSM are analyzed through experiments. This model also enables various services to be described, discovered and composed from multiple dimensions and abstraction levels with normal form and integrity guarantees. Some extensions and applications of the P-RSM are introduced.

A Periodic State Space Model to Monthly Long-term Temperature Data

This work presents a periodic state space model to model monthly temperature data. Additionally, some issues are discussed, as the parameter estimation or the Kalman filter recursions adapted to a periodic model. This framework is applied to monthly long-term temperature time series of Lisbon.

On the Set of Locally Convex Topologies Compatible with a Given Topology on a Vector Space: Cardinality Aspects

For a topological vector space (X, τ ), we consider the family LCT (X, τ ) of all locally convex topologies defined on X, which give rise to the same continuous linear functionals as the original topology τ . We prove that for an infinite-dimensional reflexive Banach space (X, τ ), the cardinality of LCT (X, τ ) is at least c.

Learning techniques for automatic email message tagging

A organização automática de mensagens de correio electrónico é um desafio actual na área da aprendizagem automática. O número excessivo de mensagens afecta cada vez mais utilizadores, especialmente os que usam o correio electrónico como ferramenta de comunicação e trabalho. Esta tese aborda o problema da organização automática de mensagens de correio electrónico propondo uma solução que tem como objectivo a etiquetagem automática de mensagens. A etiquetagem automática é feita com recurso às pastas de correio electrónico anteriormente criadas pelos utilizadores, tratando-as como etiquetas, e à sugestão de múltiplas etiquetas para cada mensagem (top-N). São estudadas várias técnicas de aprendizagem e os vários campos que compõe uma mensagem de correio electrónico são analisados de forma a determinar a sua adequação como elementos de classificação. O foco deste trabalho recai sobre os campos textuais (o assunto e o corpo das mensagens), estudando-se diferentes formas de representação, selecção de características e algoritmos de classificação. É ainda efectuada a avaliação dos campos de participantes através de algoritmos de classificação que os representam usando o modelo vectorial ou como um grafo. Os vários campos são combinados para classificação utilizando a técnica de combinação de classificadores Votação por Maioria. Os testes são efectuados com um subconjunto de mensagens de correio electrónico da Enron e um conjunto de dados privados disponibilizados pelo Institute for Systems and Technologies of Information, Control and Communication (INSTICC). Estes conjuntos são analisados de forma a perceber as características dos dados. A avaliação do sistema é realizada através da percentagem de acerto dos classificadores. Os resultados obtidos apresentam melhorias significativas em comparação com os trabalhos relacionados.

Automated illustration of multimedia stories

Submitted in part fulfillment of the requirements for the degree of Master in Computer Science

Classificação automática de textos baseada em ontologias

Dissertação de Mestrado em Engenharia Informática

A framework for supporting knowledge representation – an ontological based approach

Dissertação para obtenção do Grau de Mestre em Engenharia Electrotécnica e de Computadores

Semantic enrichment of knowledge sources supported by domain ontologies

This thesis introduces a novel conceptual framework to support the creation of knowledge representations based on enriched Semantic Vectors, using the classical vector space model approach extended with ontological support. One of the primary research challenges addressed here relates to the process of formalization and representation of document contents, where most existing approaches are limited and only take into account the explicit, word-based information in the document. This research explores how traditional knowledge representations can be enriched through incorporation of implicit information derived from the complex relationships (semantic associations) modelled by domain ontologies with the addition of information presented in documents. The relevant achievements pursued by this thesis are the following: (i) conceptualization of a model that enables the semantic enrichment of knowledge sources supported by domain experts; (ii) development of a method for extending the traditional vector space, using domain ontologies; (iii) development of a method to support ontology learning, based on the discovery of new ontological relations expressed in non-structured information sources; (iv) development of a process to evaluate the semantic enrichment; (v) implementation of a proof-of-concept, named SENSE (Semantic Enrichment kNowledge SourcEs), which enables to validate the ideas established under the scope of this thesis; (vi) publication of several scientific articles and the support to 4 master dissertations carried out by the department of Electrical and Computer Engineering from FCT/UNL. It is worth mentioning that the work developed under the semantic referential covered by this thesis has reused relevant achievements within the scope of research European projects, in order to address approaches which are considered scientifically sound and coherent and avoid “reinventing the wheel”.

English-Malayalam Cross-Lingual Information Retrieval – an Experience

This paper describes about an English-Malayalam Cross-Lingual Information Retrieval system. The system retrieves Malayalam documents in response to query given in English or Malayalam. Thus monolingual information retrieval is also supported in this system. Malayalam is one of the most prominent regional languages of Indian subcontinent. It is spoken by more than 37 million people and is the native language of Kerala state in India. Since we neither had any full-fledged online bilingual dictionary nor any parallel corpora to build the statistical lexicon, we used a bilingual dictionary developed in house for translation. Other language specific resources like Malayalam stemmer, Malayalam morphological root analyzer etc developed in house were used in this work

TU Graz: Course: 707.000 Web Science and Web Technology: Lecture 10: Text Mining

This class introduces basics of web mining and information retrieval including, for example, an introduction to the Vector Space Model and Text Mining. Guest Lecturer: Dr. Michael Granitzer Optional: Modeling the Internet and the Web: Probabilistic Methods and Algorithms, Pierre Baldi, Paolo Frasconi, Padhraic Smyth, Wiley, 2003 (Chapter 4, Text Analysis)

Hierarchical clustering based web service discovery

This paper presents a hierarchical clustering method for semantic Web service discovery. This method aims to improve the accuracy and efficiency of the traditional service discovery using vector space model. The Web service is converted into a standard vector format through the Web service description document. With the help of WordNet, a semantic analysis is conducted to reduce the dimension of the term vector and to make semantic expansion to meet the user’s service request. The process and algorithm of hierarchical clustering based semantic Web service discovery is discussed. Validation is carried out on the dataset.

Automatic Indexing and Information Visualization: A Study Based on Paraconsistent Logic

This paper reports a research to evaluate the potential and the effects of use of annotated Paraconsistent logic in automatic indexing. This logic attempts to deal with contradictions, concerned with studying and developing inconsistency-tolerant systems of logic. This logic, being flexible and containing logical states that go beyond the dichotomies yes and no, permits to advance the hypothesis that the results of indexing could be better than those obtained by traditional methods. Interactions between different disciplines, as information retrieval, automatic indexing, information visualization, and nonclassical logics were considered in this research. From the methodological point of view, an algorithm for treatment of uncertainty and imprecision, developed under the Paraconsistent logic, was used to modify the values of the weights assigned to indexing terms of the text collections. The tests were performed on an information visualization system named Projection Explorer (PEx), created at Institute of Mathematics and Computer Science (ICMC - USP Sao Carlos), with available source code. PEx uses traditional vector space model to represent documents of a collection. The results were evaluated by criteria built in the information visualization system itself, and demonstrated measurable gains in the quality of the displays, confirming the hypothesis that the use of the para-analyser under the conditions of the experiment has the ability to generate more effective clusters of similar documents. This is a point that draws attention, since the constitution of more significant clusters can be used to enhance information indexing and retrieval. It can be argued that the adoption of non-dichotomous (non-exclusive) parameters provides new possibilities to relate similar information.

A novel XML document structure comparison framework based-on sub-tree commonalities and label semantics

XML similarity evaluation has become a central issue in the database and information communities, its applications ranging over document clustering, version control, data integration and ranked retrieval. Various algorithms for comparing hierarchically structured data, XML documents in particular, have been proposed in the literature. Most of them make use of techniques for finding the edit distance between tree structures, XML documents being commonly modeled as Ordered Labeled Trees. Yet, a thorough investigation of current approaches led us to identify several similarity aspects, i.e., sub-tree related structural and semantic similarities, which are not sufficiently addressed while comparing XML documents. In this paper, we provide an integrated and fine-grained comparison framework to deal with both structural and semantic similarities in XML documents (detecting the occurrences and repetitions of structurally and semantically similar sub-trees), and to allow the end-user to adjust the comparison process according to her requirements. Our framework consists of four main modules for (i) discovering the structural commonalities between sub-trees, (ii) identifying sub-tree semantic resemblances, (iii) computing tree-based edit operations costs, and (iv) computing tree edit distance. Experimental results demonstrate higher comparison accuracy with respect to alternative methods, while timing experiments reflect the impact of semantic similarity on overall system performance.