by Law Hing Man, Hudson. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1990. / Bibliography: leaves 109-112. / Acknowledgements / Abstract / Chapter Chapter I. --- Introduction --- p.1 / Chapter 1.1 --- Overview of the thesis --- p.2 / Chapter 1.2 --- Organization of the thesis --- p.7 / Chapter Chapter II. --- Artificial Neural Networks --- p.9 / Chapter 2.1 --- Architectures of Artificial Neural Networks --- p.10 / Chapter 2.1.1 --- Neuron Models --- p.10 / Chapter 2.1.2 --- Network Models --- p.12 / Chapter 2.2 --- Learning algorithms --- p.13 / Chapter Chapter III. --- From Logic Neuron to Non-Overlapped Trees --- p.15 / Chapter 3.1 --- Deterministic Logic Neuron (DLN) --- p.15 / Chapter 3.2 --- Probabilistic Logic Neuron (PLN) --- p.20 / Chapter 3.2.1 --- Well-behaved learning of orthogonal patterns in PLN network --- p.23 / Chapter 3.2.2 --- Well-behaved learning algorithm for non-orthogonal patterns --- p.23 / Chapter 3.3 --- Non-Overlapped Trees --- p.28 / Chapter 3.3.1 --- Homogeneous learning algorithm --- p.30 / Chapter 3.3.2 --- An external comparator --- p.34 / Chapter 3.3.3 --- Problems solved by NOTPLN --- p.35 / Chapter Chapter IV. --- Properties of NOTPLN --- p.37 / Chapter 4.1 --- Noise Insensitivity --- p.37 / Chapter 4.1.1 --- Noise insensitivity with one bit noise --- p.38 / Chapter 4.1.2 --- Noise insensitivity under different noise distributions --- p.40 / Chapter 4.2 --- Functionality --- p.46 / Chapter 4.3 --- Capacity --- p.49 / Chapter 4.4 --- Distributed representation --- p.50 / Chapter 4.5 --- Generalization --- p.51 / Chapter 4.5.1 --- Text-to-Phoneme Problem --- p.52 / Chapter 4.5.2 --- Automobile Learning --- p.53 / Chapter Chapter V. --- Learning Algorithms --- p.54 / Chapter 5.1 --- Presentation methods --- p.54 / Chapter 5.2 --- Learning algorithms --- p.56 / Chapter 5.2.1 --- Heterogeneous algorithm --- p.57 / Chapter 5.2.2 --- Conflict reduction agorithm --- p.61 / Chapter 5.3 --- Side effects of learning algorithms --- p.68 / Chapter 5.3.1 --- Existence of Side Effects --- p.68 / Chapter 5.3.2 --- Removal of Side Effects --- p.69 / Chapter Chapter VI. --- Practical Considerations --- p.71 / Chapter 6.1 --- Input size constraint --- p.71 / Chapter 6.2 --- Limitations of functionality --- p.72 / Chapter 6.3 --- Thermometer code --- p.72 / Chapter 6.4 --- Output definitions --- p.73 / Chapter 6.5 --- More trees for one bit --- p.74 / Chapter 6.6 --- Repeated recall --- p.75 / Chapter Chapter VII. --- Implementation and Simulations --- p.78 / Chapter 7.1 --- Implementation --- p.78 / Chapter 7.2 --- Simulations --- p.81 / Chapter 7.2.1 --- Parity learning --- p.81 / Chapter 7.2.2 --- Performance of learning algorithms under different hamming distances --- p.82 / Chapter 7.2.3 --- Performance of learning algorithms with different output size --- p.83 / Chapter 7.2.4 --- Numerals recognition and noise insensitivity --- p.84 / Chapter 7.2.5 --- Automobile learning and generalization --- p.86 / Chapter Chapter VIII. --- Spoken Numerals Recognition System based on NOTPLN --- p.89 / Chapter 8.1 --- End-point detection --- p.90 / Chapter 8.2 --- Linear Predictive Analysis --- p.91 / Chapter 8.3 --- Formant Frequency Extraction --- p.93 / Chapter 8.4 --- Coding --- p.95 / Chapter 8.5 --- Results and discussion --- p.96 / Chapter Chapter IX. --- Concluding Remarks --- p.97 / Chapter 9.1 --- Revisit of the contributions of the thesis --- p.97 / Chapter 9.2 --- Further researches --- p.99 / Chapter Appendix A --- Equation for calculating the probability of random selection --- p.102 / Chapter Appendix B --- Training sets with different hamming distances --- p.103 / Chapter Appendix C --- Set of numerals with their associated binary values --- p.107 / References --- p.109

by Wong, Yin Pong Anthony. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (leaves 101-107). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Constraint Satisfaction Problems --- p.2 / Chapter 1.2 --- Solution Stability in Dynamic Constraint Satisfaction Problems --- p.3 / Chapter 1.3 --- Motivation of the Research --- p.5 / Chapter 1.4 --- Overview of the Thesis --- p.5 / Chapter 2 --- Related Work --- p.7 / Chapter 2.1 --- Complete Search Algorithms --- p.7 / Chapter 2.1.1 --- DnAC-4 --- p.8 / Chapter 2.1.2 --- ac --- p.9 / Chapter 2.1.3 --- DnAC-6 --- p.9 / Chapter 2.2 --- Algorithms for Stability --- p.10 / Chapter 2.2.1 --- Bellicha --- p.10 / Chapter 2.2.2 --- Dynamic Dynamic Backtracking --- p.11 / Chapter 2.2.3 --- Wallace and Freuder --- p.12 / Chapter 2.2.4 --- Unimodular Probing --- p.13 / Chapter 2.2.5 --- Train Rescheduling --- p.14 / Chapter 2.3 --- Constrained Optimization Algorithms --- p.14 / Chapter 2.3.1 --- Guided Local Search --- p.14 / Chapter 2.3.2 --- Anytime CSA with Iterative Deepening --- p.15 / Chapter 2.4 --- A Real-life Application --- p.16 / Chapter 3 --- Background --- p.17 / Chapter 3.1 --- Fuzzy Constraint Satisfaction Problems --- p.17 / Chapter 3.2 --- Fuzzy GENET --- p.19 / Chapter 3.2.1 --- Network Architecture --- p.19 / Chapter 3.2.2 --- Convergence Procedure --- p.21 / Chapter 3.3 --- Deficiency in Fuzzy GENET --- p.24 / Chapter 3.4 --- Rectification of Fuzzy GENET --- p.26 / Chapter 4 --- Using Fuzzy GENET for Solving Stability Problems --- p.30 / Chapter 4.1 --- Modelling Stability Problems as FCSPs --- p.30 / Chapter 4.2 --- Extending Fuzzy GENET for Solving Stability Problems --- p.36 / Chapter 4.3 --- Experiments --- p.38 / Chapter 4.3.1 --- Dynamic CSP Generation --- p.39 / Chapter 4.3.2 --- Problems Using Hamming Distance Function --- p.41 / Chapter 4.3.2.1 --- Variation in Number of Variables --- p.42 / Chapter 4.3.2.2 --- Variation in Domain Size --- p.45 / Chapter 4.3.2.3 --- Variation in Density and Tightness --- p.47 / Chapter 4.3.3 --- Comparison in Using Different Thresholds --- p.47 / Chapter 4.3.4 --- Problems Using Manhattan Distance Function --- p.50 / Chapter 5 --- Enhancement of the Modelling Scheme --- p.56 / Chapter 5.1 --- Distance Bound --- p.56 / Chapter 5.2 --- Enhancement of Convergence Procedure --- p.57 / Chapter 5.3 --- Comparison with Optimal Solutions --- p.60 / Chapter 5.4 --- Comparison with Fuzzy GENET(dcsp) --- p.64 / Chapter 5.4.1 --- Medium-sized Problems --- p.64 / Chapter 5.4.2 --- The 150-10-15-15 Problem --- p.67 / Chapter 5.4.3 --- Variation in Density and Tightness --- p.73 / Chapter 5.4.4 --- Variation in Domain Size --- p.76 / Chapter 5.5 --- Analysis of Fuzzy GENET(dcsp2) --- p.94 / Chapter 6 --- Conclusion --- p.98 / Chapter 6.1 --- Contributions --- p.98 / Chapter 6.2 --- Future Work --- p.99 / Bibliography --- p.101

Law Yat Chiu. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2002. / Includes bibliographical references (leaves 85-89). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Related Work --- p.4 / Chapter 2.1 --- Equivalence of CSPs --- p.4 / Chapter 2.2 --- Dual Viewpoint --- p.4 / Chapter 2.3 --- CSP Reformulation --- p.5 / Chapter 2.4 --- Multiple Modeling --- p.5 / Chapter 2.5 --- Redundant Modeling --- p.6 / Chapter 2.6 --- Minimal Combined Model --- p.6 / Chapter 2.7 --- Permutation CSPs and Channeling Constraints --- p.6 / Chapter 3 --- Background --- p.8 / Chapter 3.1 --- From Viewpoints to CSP Models --- p.8 / Chapter 3.2 --- Constraint Satisfaction Techniques --- p.10 / Chapter 3.2.1 --- Backtracking Search --- p.11 / Chapter 3.2.2 --- Consistency Techniques and Constraint Propagation --- p.12 / Chapter 3.2.3 --- Incorporating Consistency Techniques into Backtracking Search --- p.18 / Chapter 4 --- Model Induction --- p.21 / Chapter 4.1 --- Channeling Constraints --- p.21 / Chapter 4.2 --- Induced Models --- p.22 / Chapter 4.3 --- Properties --- p.30 / Chapter 5 --- Exploiting Redundancy from Model Induction --- p.35 / Chapter 5.1 --- Combining Redundant Models --- p.35 / Chapter 5.1.1 --- Model Intersection --- p.36 / Chapter 5.1.2 --- Model Channeling --- p.38 / Chapter 5.2 --- Three New Forms of Model Redundancy --- p.39 / Chapter 5.3 --- Experiments --- p.42 / Chapter 5.3.1 --- Langford's Problem --- p.44 / Chapter 5.3.2 --- Random Permutation CSPs --- p.53 / Chapter 5.3.3 --- Golomb Rulers --- p.72 / Chapter 5.3.4 --- Circular Golomb Rulers --- p.74 / Chapter 5.3.5 --- All-Interval Series Problem --- p.78 / Chapter 6 --- Concluding Remarks --- p.82 / Chapter 6.1 --- Contributions --- p.82 / Chapter 6.2 --- Future Work --- p.83

by Benson, Ng Hin Kwong. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (leaves 111-117). / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.9 / Chapter 1.1 --- Our approach --- p.11 / Chapter 1.2 --- Organization of the thesis --- p.12 / Chapter 2 --- Logic Programming --- p.13 / Chapter 2.1 --- Logic programming in Horn clauses --- p.14 / Chapter 2.1.1 --- Problem with incomplete information --- p.15 / Chapter 2.1.2 --- Problem with inconsistent information --- p.15 / Chapter 2.1.3 --- Problem with indefinite information --- p.16 / Chapter 2.2 --- Logic programming in non-Horn clauses --- p.16 / Chapter 2.2.1 --- Reasoning under incomplete information --- p.17 / Chapter 2.2.2 --- Reasoning under inconsistent information --- p.17 / Chapter 2.2.3 --- Reasoning under indefinite information --- p.20 / Chapter 2.3 --- "Coexistence of incomplete, inconsistent and indefinite information" --- p.21 / Chapter 2.4 --- Stable semantics --- p.22 / Chapter 2.5 --- Well-founded semantics --- p.23 / Chapter 2.6 --- Chapter summary --- p.25 / Chapter 3 --- Argumentation --- p.26 / Chapter 3.1 --- Toulmin's informal argumentation model --- p.27 / Chapter 3.2 --- Rescher's formal argumentation model --- p.28 / Chapter 3.3 --- Argumentation in AI research --- p.30 / Chapter 3.3.1 --- Poole's Logical Framework for Default Reasoning --- p.30 / Chapter 3.3.2 --- Inheritance Reasoning Framework of Touretzky et. al --- p.31 / Chapter 3.3.3 --- Pollock's Theory of Defeasible Reasoning --- p.32 / Chapter 3.3.4 --- Dung's Abstract Argumentation Framework --- p.33 / Chapter 3.3.5 --- Lin and Shoham's Argument System --- p.35 / Chapter 3.3.6 --- Vreeswijk's Abstract Argumentation --- p.35 / Chapter 3.3.7 --- Kowalski and Toni's Uniform Argumentation --- p.36 / Chapter 3.3.8 --- John Fox's Qualitative Argumentation --- p.37 / Chapter 3.3.9 --- Thomas Gordon's Pleading Games --- p.38 / Chapter 3.3.10 --- Chris Reed's Persuasive Dialogue --- p.39 / Chapter 3.3.11 --- Ronald Loui's Argument Game --- p.39 / Chapter 3.3.12 --- "Verheij's Reason-Based, Logics and CumulA" --- p.40 / Chapter 3.3.13 --- Prakken's Defeasible Argumentation --- p.40 / Chapter 3.3.14 --- Summary of existing frameworks --- p.41 / Chapter 3.4 --- Chapter summary --- p.42 / Chapter 4 --- Disjunctive Argumentation Semantics I --- p.46 / Chapter 4.1 --- Background --- p.47 / Chapter 4.2 --- Definition --- p.48 / Chapter 4.3 --- Conflicts within a KBS --- p.52 / Chapter 4.4 --- Conflicts between KBSs --- p.54 / Chapter 4.4.1 --- Credulous View --- p.56 / Chapter 4.4.2 --- Skeptical View --- p.57 / Chapter 4.4.3 --- Generalized Skeptical View --- p.58 / Chapter 4.5 --- Semantics --- p.60 / Chapter 4.6 --- Dialectical proof theory --- p.61 / Chapter 4.7 --- Relation to existing framework --- p.61 / Chapter 4.8 --- Issue on paraconsistency --- p.63 / Chapter 4.9 --- An illustrative example --- p.63 / Chapter 4.10 --- Chapter summary --- p.65 / Chapter 5 --- Disjunctive Argumentation Semantics II --- p.67 / Chapter 5.1 --- Background --- p.68 / Chapter 5.2 --- Definition --- p.70 / Chapter 5.2.1 --- Rules --- p.70 / Chapter 5.2.2 --- Splits --- p.71 / Chapter 5.3 --- Conflicts --- p.74 / Chapter 5.3.1 --- Undercut conflicts --- p.75 / Chapter 5.3.2 --- Rebuttal conflicts --- p.76 / Chapter 5.3.3 --- Thinning conflicts --- p.78 / Chapter 5.4 --- Semantics --- p.80 / Chapter 5.5 --- Relation to existing frameworks --- p.81 / Chapter 5.6 --- Issue on paraconsistency --- p.82 / Chapter 5.7 --- An illustrative example --- p.83 / Chapter 5.8 --- Chapter summary --- p.85 / Chapter 6 --- Evaluation --- p.86 / Chapter 6.1 --- Introduction --- p.86 / Chapter 6.2 --- Methodology --- p.87 / Chapter 6.3 --- DAS I --- p.88 / Chapter 6.3.1 --- Inoue's Benchmark problems --- p.88 / Chapter 6.3.2 --- Sherlock Holmes' problems --- p.96 / Chapter 6.4 --- DAS II --- p.100 / Chapter 6.4.1 --- Inoue's benchmark problems --- p.100 / Chapter 6.4.2 --- Sherlock Holmes' problem --- p.103 / Chapter 6.5 --- Analysis --- p.103 / Chapter 6.5.1 --- Possible extension --- p.104 / Chapter 6.6 --- Chapter summary --- p.106 / Chapter 7 --- Conclusion --- p.108 / Chapter 7.0.1 --- Possible extension of the present work --- p.109 / Bibliography --- p.117 / Chapter A --- First Oreder Logic (FOL) --- p.118 / Chapter B --- DAS-I Proof --- p.121 / Chapter B.1 --- Monotone proof --- p.121 / Chapter B.2 --- Soundness proof --- p.122 / Chapter B.3 --- Completeness proof --- p.123 / Chapter C --- Sherlock Holmes' Silver Blaze Excerpts --- p.125 / Chapter C.1 --- Double life --- p.125 / Chapter C.2 --- Poison stable boy --- p.125

<p> Within the realm of computational story generation sits Minstrel, a decades old system which was once used to explore the idea that, under the correct conditions, novel stories can be generated by taking an existing story and replacing some of its elements with similar ones found in a different story. This concept would eventually fall within the bounds of a strategy known as Case-Based Reasoning (CBR), in which problems are solved by recalling solutions to past problems (the cases), and mutating the recalled cases in order to create an appropriate solution to the current problem. This dissertation uses a rational reconstruction of Minstrel called Minstrel Remixed, a handful of upgraded variants of Minstrel Remixed, and a pair of similar but unrelated storytelling systems, to explore various characteristics of Minstrel-style storytelling systems. </p><p> In the first part of this dissertation I define the class of storytelling systems that are similar to Minstrel. This definition allows me to compare the features of these systems and discuss the various strengths and weaknesses of the variants. Furthermore, I briefly describe the rational reconstruction of Minstrel and then provide a detailed overview of the inner workings of the resulting system, Minstrel Remixed. </p><p> Once Minstrel Remixed was complete, I chose to upgrade it in order to explore the set of stories that it could produced and ways to alter or reconfigure the system with the goal of intentionally influencing the set of possible outputs. This investigation resulted in two new storytelling systems called Conspiracy Forever and Problem Planets. The second portion of this dissertation discusses these systems as well as a number of discoveries about the strengths and weaknesses of Minstrel Style Storytelling Systems in general. More specifically, I discuss that, 1) a human reader's capacity for creating patterns out of an assortment of statements is incredibly useful and output should be crafted to use this potential, 2) Minstrel-Style Storytelling tends to be amnesiac and do a poor job of creating long stories that remain cohesive, and 3) the domain that a storytelling system is working from is incredibly important and must be well engineered. I continue by discussing the methods that I discovered for cleaning up and maintaining a domain and conclude with a section covering interviews with other storytelling system creators about the strengths and weaknesses of their systems in light of my findings about Minstrel Remixed. </p><p> In the final portion of this document I create a framework of six interrelated attributes of stories (length, coherence, creativity, complexity, contextuality, and consolidation,) and use this along with the learning discussed in the first two portions of the dissertation to discuss the strengths and weaknesses of this class of CBR systems when applied to both static story generation and interactive storytelling. I discuss the finding that these systems seem to have some amount of power and although they can be tweaked to produce for example, longer or more consolidated stories, these improvements always come along with a reduction in complexity, coherence, or one of the other attributes. Further discussion of the output power of this class of storytelling systems revolves around the primary limiting factor to their potential, namely the fact that they have no understanding of the symbols and patterns that they are manipulating. Finally, I introduce a number of strategies that I found to be fruitful for increasing the 'output power' of the system and getting around the lack of commonsense reasoning, chiefly improving the domain and adding new subsystems.</p><p>

<p> The overarching goal of my thesis work was to utilize the combination of mathematical and experimental models towards an effort to resolve chemotherapy-induced peripheral neuropathy (CIPN), one of the most common adverse effects of cancer chemotherapy. In chapter two, we have developed quantitative-structure toxicity relationship (QSTR) models using machine learning algorithms that enable the prediction of peripheral neuropathy incidence solely from a chemicals molecular structure. The QSTR models enable the prediction of clinical neurotoxicity, which could be potentially useful in early drug discovery to screen out compounds that are highly neurotoxic and identify safer drug candidates to move forward into further development. The QSTR model was used to suggest modifications to the molecular structure of bortezomib that may reduce the number of patients who develop peripheral neuropathy from bortezomib therapy. In the third chapter, we conducted a network-based comparative systems pharmacology analysis of proteasome inhibitions. The concept behind this work was to use <i>in silico</i> pharmacological interaction networks to elucidate the neurotoxic differences between bortezomib and carfilzomib. Our theoretical results suggested the importance of the unfolded protein response in bortezomib neurotoxicity and that the mechanisms of neurotoxicity by proteasome inhibitors closely relate to the pathogenesis of Guillian-Barr&eacute; syndrome caused by the Epstein-Barr virus. In chapter four we have written a review article to introduce the concept of Boolean network modeling in systems pharmacology. Due to the lack of knowledge about parameter values that govern the cellular dynamic processes involved in peripheral nerve damage, the development of a quantitative systems pharmacology model would not be feasible. Therefore, in chapter five, we developed a Boolean network-based systems pharmacology model of intracellular signaling and gene regulation in peripheral neurons. The model was used to simulate the neurotoxic effects of bortezomib and to identify potential treatment strategies for proteasome-inhibitor induced peripheral neuropathy. A novel combinatorial treatment strategy was identified that consists of a TNF? inhibitor, NMDA receptor antagonist, and reactive oxygen species inhibitor. Our subsequent goals were aimed towards translating this finding with the endeavor to hopefully one-day impact human health. Initially we had proposed to use three separate agents for each of these targets, however the clinical administration of three agents to prevent the neurotoxicity of one is likely unfeasible. We then came across a synthetic cannabinoid derivative, dexanabinol, that promiscuously inhibits all three of these targets and was previously developed for its intended use to treat traumatic brain injury. We believe that this drug candidate was worth investigating due to the overlapping pharmacological activity with suggested targets from network analyses, previously established favorable safety profile in humans, notable <i>in vitro/vivo</i> neuroprotective properties, and rising popularity for the therapeutic potential of cannabinoids to treat CIPN. In chapter six we assessed the efficacy of dexanabinol for preventing the neurotoxic effects of bortezomib in various experimental models. Due to the limited translatability of 2D cell culture techniques, we investigated the pharmacodynamics of dexanabinol using a microphysiological model of the peripheral nerve. Bortezomib caused a reduction in electrophysiological endpoints, which were partially restored by dexanabinol. In chapter 7 we evaluated the possible interaction of dexanabinol on the anti-cancer effects of bortezomib. We observed no significant differences in tumor volume between bortezomib alone and in combination with dexanabinol in a multiple myeloma mouse model. Lastly, we are currently investigating the efficacy of dexanabinol in well-established rat model of bortezomib-induced peripheral neuropathy. We believe that positive results would warrant a clinical trial. In conclusion, the statistical and mechanistic models of peripheral neuropathy that were developed could be used to reduce the overall burden of CIPN through the design of safer chemotherapeutics and discovery of novel neuroprotective treatment strategies.</p><p>

<p> This document proposes new methods for training multi-layer and deep spiking neural networks (SNNs), specifically, spiking convolutional neural networks (CNNs). Training a multi-layer spiking network poses difficulties because the output spikes do not have derivatives and the commonly used backpropagation method for non-spiking networks is not easily applied. Our methods use novel versions of the brain-like, local learning rule named spike-timing-dependent plasticity (STDP) that incorporates supervised and unsupervised components. Our method starts with conventional learning methods and converts them to spatio-temporally local rules suited for SNNs. </p><p> The training uses two components for unsupervised feature extraction and supervised classification. The first component refers to new STDP rules for spike-based representation learning that trains convolutional filters and initial representations. The second introduces new STDP-based supervised learning rules for spike pattern classification via an approximation to gradient descent by combining the STDP and anti-STDP rules. Specifically, the STDP-based supervised learning model approximates gradient descent by using temporally local STDP rules. Stacking these components implements a novel sparse, spiking deep learning model. Our spiking deep learning model is categorized as a variation of spiking CNNs of integrate-and-fire (IF) neurons with performance comparable with the state-of-the-art deep SNNs. The experimental results show the success of the proposed model for image classification. Our network architecture is the only spiking CNN which provides bio-inspired STDP rules in a hierarchy of feature extraction and classification in an entirely spike-based framework.</p><p>

Made available in DSpace on 2019-03-30T00:01:38Z (GMT). No. of bitstreams: 0 Previous issue date: 2016-11-28 / An emerging Natural Language Processing (NLP) research field proposes Open Information Extraction Systems (Open IE systems) which the main feature is do not need predefined semantic relations for text extraction and instead of this aims at generic standards to extract any domain-independent information. Following this paradigm, this work introduces RePort - an Open Information Extraction System for Portuguese Language, which is designed to scale massive data bases and extract any kind of verb-mediated relationships from Portuguese plain text files. The work¿s other contributions are as follows: a Golden Standard relation-labeled sentences in Portuguese; a lexical database of verbal relations generated from CETENFolha corpus; and generic methods for the creation and evolution of this lexical database of verbal relations via corpora or web queries. Experimental evaluations in English-Portuguese bilingual corpus show the need for linguistic knowledge to adapt the correlated system ¿ ReVerb from English to Portuguese. Based on a second analysis, an automatic evaluation of RePort achieves best results using the extended lexical database of verbal relations and is near to the state-of-the art, when considered only the extraction¿s verbal relations. Finally, it is important to point out the importance of the RePort system, and of the other contributions and analyzes show here aim at evolution of the Open IE system area for Portuguese Language. Keywords: Computational Linguistics, Artificial Intelligence, Natural Language Processing, Information Extraction, Open Information Extraction System, Portuguese Language. / Um campo emergente de pesquisa em Processamento e Linguagem Natural (PLN) propõe Sistemas de Extração de Informações Aberta (em inglês - Open Information Extraction Systems - Open IE systems, em inglês) que têm como a principal característica não necessitar de definição a priori dos tipos de relações semânticas a serem extraídas de textos, visando padrões genéricos para a extração de quaisquer informações independente de domínio. Seguindo este paradigma, este trabalho apresenta o RePort ¿ um Sistema de Extração de Informações Aberta para Língua Portuguesa, projetado para escalar bases massivas de dados e extrair de quaisquer tipos de relações mediadas por verbo a partir de documentos textuais em português. Como contribuições secundárias deste trabalho têm-se um Golden Standard composto dos textos e suas respectivas relações semânticas anotadas; um léxico de relações verbais gerado a partir do corpus CETENFolha; e métodos genéricos para criação e evolução do léxico de relações verbais a partir de corpora ou consultas na Web. Avaliações experimentais em corpus bilíngue inglês-português evidenciou a necessidade de conhecimento linguístico para adaptar o sistema correlato em língua inglesa ¿ ReVerb. Em uma segunda análise, avaliações automáticas do RePort apontou que este obteve seus melhores resultados utilizando o léxico de relações verbais ampliado, próximo ao estado da arte, quando considerada apenas a extração de relações verbais. Por fim, cumpre salientar a importância do sistema RePort, e das demais contribuições e análises aqui apresentadas para evolução da área de Open IE system para o português. Palavras-chave: Linguística Computacional, Inteligência Artificial, Processamento de Linguagem Natural, Extração de Informações, Sistemas de Extração de Informações Aberta, Língua Portuguesa.

Não disponibilizado pelo autor. / Not available.

<p> With the rising use of autonomous agents within robotic and software settings, agents may be required to cooperate in teams while having little or no information regarding the capabilities of their teammates. In these ad hoc settings, teams must collaborate on the fly, having no prior opportunity for coordination. Prior research in this area commonly either assumes that communication between agents is impossible given their heterogeneous design or has left communication as an open problem. Typically, to accurately predict a teammate's behavior at a future point in time, ad hoc agents leverage models learned from past experience and attempt to infer a teammate's intended strategy through observing its current course of action. However, these approaches can fail to arrive at accurate policy predictions, leaving the coordinating agent uncertain and unable to adapt to its teammates' plans. We introduce the problem of communicating minimal sets of teammate policies in order to provide information for collaboration in such ad hoc environments. We demonstrate how an agent may determine what information it should solicit from its peers but further illustrate how optimal solutions to such a problem have intractable computational requirements. Nonetheless, through the characterization of this difficulty, we identify strategies that permit approximate or heuristic approaches, allowing the practical application of this capacity in ad hoc teams.</p><p>