Na última década métodos sintáticos vêm sendo bastante empregados para reconhecimento de padrões em imagens. Uma revisão sistemática sobre o assunto indicou que há trabalhos sobre a utilização de gramáticas para reconhecimento de objetos específicos em imagens, reconhecimento de texturas, construção de objetos, segmentação de imagens, mudança de escala de imagens e reconhecimento de layouts de páginas e documentos. Foi percebido que apenas alguns dos trabalhos analisados apresentavam métodos de inferência gramatical. Percebida essa lacuna, este trabalho teve como objetivo avaliar a utilização de gramáticas estocásticas inferidas a partir de amostras de treinamento que continham as estruturas de quadtrees das imagens no problema de classificação de imagens. Para atingir tal objetivo foram utilizados e adaptados dois algoritmos de inferência gramatical e um algoritmo de estimação de probabilidades responsável por transformar as gramáticas livres de contexto inferidas em gramáticas livres de contexto estocásticas. Os resultados obtidos mostram que as técnicas propostas podem ser utilizadas na classificação de figuras geométricas. Entretanto, embora o desempenho com quadtrees não tenha sido excelente, ficou claro que o uso de gramáticas pode ser uma abordagem interessante para a classificação de imagens. / In the last decade syntactic methods have been widely used for pattern recognition in images. A systematic review on this subject indicated that there are works about the usage of grammars for recognizing specific objects in images, recognition of textures, construction of objects, image segmentation, image scaling, and recognition of layouts in documents. It was noticed that only some of the analyzed studies showed methods of grammatical inference. Perceived this gap, this study aimed to evaluate the usage of stochastic grammars inferred from training samples containing structures of quadtrees of images in image classification problem. To achieve this goal it was used and adapted two algorithms for grammatical inference and one algorithm of probability estimation responsible for transforming the inferred context-free grammars in stochastic context-free grammars. The results show that the proposed techniques can be used in the classification of geometric figures. However, although the performance with quadtrees has not been excellent, it was clear that the use of grammars can be an interesting approach for image classification.

本論文擬論述臺語述補式結構之語法與語意特性。述補式複合結構由兩個動詞或一個動詞及一個表示此一動詞之時帽組合成。由於臺語文獻中缺乏此類結構之研究，我們瀏灠了漢語中有關術補式複合動詞之研究，並採用張郇慧教授以論旨結構為本之分析模式。此結構之形成與論旨結構及句法上之及物性密切相關。除了此結構之形成外，我們亦討論其句法及語意系統與論旨結構之關聯。我們希望能藉此研究對此結構提供清楚而基本之概念。 / The previous studies on Taiwanese focus on the phonology and written forms, but few on other fields.In the thesis, we adopt a widely usde structure: Verb-Complement Construction to explore the syntax and semantics of the construction.Verb-Complement Constructions consist of two verbs or one verb plus a functional item which specifies verbal of the first verb. Due to the lack of previous studies on the topic, we first review several analyses on Mandarin Verb-Complement Compounds and adopt Chang's (1994) argument-structure-based analysis as the research approach. We propose that two factors contribute to the formation of this construction: syntactic transitivity and argument structure.The argument structure is divided into two categories: full-fledged and degenerated. Verb-complement Constructions, syntactically speaking, are formed by the combinations of transitive-transitive, intransitive-intransitive, and transitive-intransitive. Verb-Complement Constructions with full-fledged argument structure are distinguished into eight types based on the five types of this construction in terms of syntactic transitivity; Verb-complement constructions with degenerated argument structure are classified into four types based on the four kinds of this construction in terms of syntactic transitivity. We argue that different syntactic structures such as middle forms, verb copying constructions, HO construction, and KA construction as well as semantics of the construction can be effectively accounted for by argument structure.By the study, we hope to provide a fundamental and clear picture on the construction.

<p>This corpus-based study examines adjectives and adjectival expressions used to describe characters in British children’s fiction. The focus is on diachronic variation, by comparing Victorian (19th-century) and contemporary (late 20th-century) children’s fiction, and on gender variation, by comparing the descriptions of female and male characters. I adopt a qualitative as well as a quantitative approach, and consider factors such as lexical diversity, adjectival density, collocation patterns, evaluative meaning, syntactic function and distribution across semantic domains. Most findings are related to a dichotomy set up between an idealistic and a realistic portrayal of characters. The study shows that an idealistic portrayal of characters is typical of the Victorian material and a realistic portrayal of characters typical of the contemporary material. Further, gender differences are much more pronounced, and reflect traditional gender role patterns more in the Victorian material than in the contemporary material. For instance, a pleasant appearance is typically described for Victorian female characters and social position for Victorian male characters. Moreover, descriptions of mental properties of Victorian female characters are conspicuously rare. Such gendered patterns are less distinct in the contemporary material, although appearance is still more extensively described for female than male characters. As regards how the qualities are attributed to characters, the descriptions of Victorian female characters were found to be the most formulaic compared to the descriptions of Victorian male, contemporary female and contemporary male characters.</p>

ここに掲載した著作物の利用に関する注意 本著作物の著作権は（社）情報処理学会に帰属します。 本著作物は著作権者である情報処理学会の許可のもとに掲載するものです。 ご利用に当たっては「著作権法」ならびに「情報処理学会倫理綱領」 に従うことをお願いいたします。 Notice for the use of this material The copyright of this material is retained by the Information Processing Society of Japan (IPSJ). This material is published on this web site with the agreement of the author (s) and the IPSJ. Please be complied with Copyright Law of Japan and the Code of Ethics of the IPSJ if any users wish to reproduce, make derivative work, distribute or make available to the public any part or whole thereof. All Rights Reserved, Copyright (C) Information Processing Society of Japan. Comments are welcome. Mail to address: 　editj<at>ipsj.or.jp, please.

This corpus-based study examines adjectives and adjectival expressions used to describe characters in British children’s fiction. The focus is on diachronic variation, by comparing Victorian (19th-century) and contemporary (late 20th-century) children’s fiction, and on gender variation, by comparing the descriptions of female and male characters. I adopt a qualitative as well as a quantitative approach, and consider factors such as lexical diversity, adjectival density, collocation patterns, evaluative meaning, syntactic function and distribution across semantic domains. Most findings are related to a dichotomy set up between an idealistic and a realistic portrayal of characters. The study shows that an idealistic portrayal of characters is typical of the Victorian material and a realistic portrayal of characters typical of the contemporary material. Further, gender differences are much more pronounced, and reflect traditional gender role patterns more in the Victorian material than in the contemporary material. For instance, a pleasant appearance is typically described for Victorian female characters and social position for Victorian male characters. Moreover, descriptions of mental properties of Victorian female characters are conspicuously rare. Such gendered patterns are less distinct in the contemporary material, although appearance is still more extensively described for female than male characters. As regards how the qualities are attributed to characters, the descriptions of Victorian female characters were found to be the most formulaic compared to the descriptions of Victorian male, contemporary female and contemporary male characters.

Η παρούσα διατριβή έχει ως σκοπό της, πρώτον, την ανάκτηση συντακτικής πληροφορίας (αναγνώριση συμπληρωμάτων ρημάτων, ανάκτηση πλαισίων υποκατηγοριοποίησης (ΠΥ) ρημάτων, αναγνώριση των ορίων και του είδους των προτάσεων) αυτόματα μέσα από ελληνικά και αγγλικά σώματα κειμένων με την χρήση ποικίλων και καινοτόμων τεχνικών μηχανικής μάθησης και, δεύτερον, την θεωρητική περιγραφή της ελληνικής σύνταξης μέσω τυπικών γλωσσολογικών φορμαλισμών, όπως η γραμματική Ενοποίησης και η γραμματική Φραστικής Δομής Οδηγούμενη από τον Κύριο Όρο. Η διατριβή κινήθηκε πάνω στους εξής καινοτόμους άξονες: 1. Η προεπεξεργασία των σωμάτων κειμένων βασίστηκε σε ελάχιστους γλωσσολογικούς πόρους για να είναι δυνατή η μεταφορά των μεθόδων σε γλώσσες φτωχές σε υποδομή. 2. Η αντιμετώπιση του θορύβου που υπεισέρχεται στα δεδομένα εξ αιτίας της χρήσης ελάχιστων πόρων πραγματοποιείται με Μονόπλευρη Δειγματοληψία. Εντοπίζονται αυτόματα παραδείγματα δεδομένων που δεν προσφέρουν στην μάθηση και αφαιρούνται. Τα τελικά δεδομένα είναι πιο καθαρά και η απόδοση της μάθησης βελτιώνεται πολύ. 3. Αποδεικνύεται η χρησιμότητα της εξαχθείσας πληροφορίας. Η χρησιμότητα των συμπληρωμάτων φαίνεται από την αύξηση της απόδοσης της διαδικασίας ανάκτησης ΠΥ με την χρήση τους. Η χρησιμότητα των εξαγόμενων ΠΥ φαίνεται από την αύξηση της απόδοσης ενός ρηχού συντακτικού αναλυτή με την χρήση τους. 4. Οι μέθοδοι εφαρμόζονται και στα Αγγλικά και στα Ελληνικά για να φανεί η μεταφερσιμότητά τους σε διαφορετικές γλώσσες και για να πραγματοποιηθεί μια ενδιαφέρουσα σχετική σύγκριση ανάμεσα στις δύο γλώσσες. Τα αποτελέσματα είναι πολύ ενθαρρυντικά, συγκρίσιμα με, και σε πολλές περιπτώσεις καλύτερα από, προσεγγίσεις που χρησιμοποιούν εξελιγμένα εργαλεία προεπεξεργασίας. / The thesis aims firstly at the acquisition of syntactic information (detection of verb complements, acquisition of verb subcategorization frames (SF), detection of the boundaries and the semantic type of clauses) automatically from Modern Greek and English text corpora with the use of various state-of-the-art and novel machine learning techniques, and, secondly, at the theoretical description of the Greek syntax through formal grammatical theories like Unification Grammar and Head-driven Phrase Structure Grammar. The thesis has been based on the following novel axes: 1. Corpus pre-processing has been limited to the use of minimum linguistic resources to ensure the portability of the presented methodologies to languages that are poorly equipped with resources. 2. Due to the low pre-processing level, a significant amount of noise appears in the data, which is dealt with One-sided Sampling. Examples that do not contribute to the learning process are detected and removed. The final data set is clean and learning performance improves significantly. 3. The importance of the acquired information is proven. The importance of complements is shown by the improvement in the performance of the SF acquisition process after the incorporation of complement information. The importance of the acquired SF lexicon is shown by its incorporation in a shallow syntactic parser and the increase of the performance of the latter. 4. The methods are applied on Modern Greek and on English to show their portability across different languages and to allow for an interesting rough comparison between the two languages. The results are very satisfactory, comparable to, and in some cases better than, approaches utilizing sophisticated resources for pre-processing.

兩種語句處理模型曾被提出以解釋閱讀句子時語法及語意資訊的互動。句法優先模型(syntax-first models)認為詞性判斷必定先於語意分析，而制約滿足模型(constraint-satisfaction models)則認為不同的資訊在語句理解的過程中會同時被處理。本研究檢驗兩種語句處理模型能否解釋語句中的詞彙歧義解困(lexical ambiguity resolution)。 許多眼動研究曾發現詞彙歧義詞的次要語義偏向效應(subordinate bias effect)，顯示語意偏向次要語義的語境能加速激發歧義詞的次要語義並且產生語義競爭。然而，語境的語法在詞彙歧義解困中扮演的角色並不清楚。因而，不同語義分屬不同詞性的詞性歧義詞便提供我們一個媒介以檢驗詞彙歧義解困中詞性及語意限制的互動。 本研究的目的有二：(一) 檢驗詞性限制能否決定中文詞性歧義詞的語義解困；(二) 檢驗中文歧義詞語義的詞性是否會影響次要語義偏向效應。實驗一我們將四種不同類型的中文非均勢同形異義詞置於語法及語意皆為次要語義偏向的句子裡；實驗二則將四種不同類型的中文非均勢同形異義詞置於語法為次要語義偏向但語意中立的句子裡。受試者閱讀句子時的眼動表現會即時被記錄。 實驗一的結果發現：(一) 四類型歧義詞的次要語義偏向效應只反映在目標詞後區域的二次閱讀指標上；(二) 就效果量而言，NV歧義詞的次要語義偏向效應在目標詞及目標詞後兩個區域都比VN歧義詞來得大。實驗二的結果則發現：(一) VN歧義詞的次要語義偏向效應從目標詞區的首次閱讀指標就出現，並且持續至目標詞及目標詞後兩個區域的二次閱讀指標；(二) 另外三類型歧義詞的次要語義偏向效應直到所有分析區域的二次閱讀指標才反映出來；(三) NV歧義詞的次要語義偏向效應比VN歧義詞出現得更晚也更不明顯。整體而言，本研究的結果顯示詞性限制並不是影響中文詞性歧義詞語義解困的唯一因素。此結果支持制約滿足模型，並反對句法優先模型的預測。 / Two primary sentence processing models have been proposed to account for the interaction between syntactic and semantic information in reading sentences: Syntax-first models assume that syntactic-category assignment must precede semantic analysis, while constraint-satisfaction models propose that information from different sources is processed and weighed at the same time during sentence comprehension. The present study examined whether these sentence processing models, which assume different contribution of syntactic category and semantic context, can explain the resolution of lexical ambiguity in sentences. Several eye movement studies have demonstrated the subordinate bias effect (SBE) for lexical-semantic ambiguous words (i.e., NN/VV homographs), indicating that a subordinate-biased semantic context can boost the activation of the subordinate meaning of ambiguous words and causes meaning competition (Duffy, Morris, & Rayner, 1988). However, the role of syntactic context in lexical ambiguity resolution is less clear. Syntactic category ambiguous words (i.e., SCA words; VN/NV homographs), whose alternative meanings differ in syntactic category (e.g., watch in English), serve as a means of examining the interaction between syntactic category and semantic constraints during lexical ambiguity resolution. The purpose of the present study was twofold: (a) to examine whether the syntactic category constraint can determine the semantic resolution of Chinese SCA words, and (b) to investigate whether syntactic category of alternative meanings of Chinese homographs can influence the SBE during lexical ambiguity resolution. Four types of Chinese biased homographs (NN, VV, VN, and NV) were embedded into syntactically and semantically subordinate-biased sentences (Experiment 1) and into syntactically subordinate-biased but semantically neutral sentences (Experiment 2). Participants’ eye movements were recorded as they read each sentence. In Experiment 1, the results showed: (1) The SBE for the four types of homographs was significant only in the second-pass reading on the post-target words. (2) Numerically, the NV homographs revealed a larger effect size of SBE than VN homographs on both target and post-target words. In Experiment 2, the results showed: (1) The SBE for VN appeared from the first-pass reading on the target words and lasted to the second-pass reading on the target and post-target words. (2) The SBE for the other types of homographs did not occur until the second-pass reading in all analyzed regions. (3) The SBE for NV occurred much later and less obviously than that for VN. In general, our findings support the constraint-satisfaction models and reject the prediction of the syntax-first models, suggesting that the syntactic category constraint is not the only factor influencing the semantic resolution of SCA words.

Depuis quelques années, les applications intégrant un module de dialogues avancés sont en plein essor. En revanche, le processus d’universalisation de ces systèmes est rapidement décourageant : ceux-ci étant naturellement dépendants de la langue pour laquelle ils ont été conçus, chaque nouveau langage à intégrer requiert son propre temps de développement. Un constat qui ne s’améliore pas en considérant que la qualité est souvent tributaire de la taille de l’ensemble d’entraînement. Ce projet cherche donc à accélérer le processus. Il rend compte de différentes méthodes permettant de générer des versions polyglottes d’un premier système fonctionnel, à l’aide de la traduction statistique. L’information afférente aux données sources est projetée afin de générer des données cibles parentes, qui diminuent d’autant le temps de développement subséquent. En ce sens, plusieurs approches ont été expérimentées et analysées. Notamment, une méthode qui regroupe les données avant de réordonner les différents candidats de traduction permet d’obtenir de bons résultats. / For a few years now, there has been an increasing number of applications allowing advanced dialog interactions with the user. However, the universalization of those systems quickly becomes painful : since they are highly dependent on the original development language, each new language to integrate requires an additionnal and significative time investment. A matter that only gets worse considering quality usually rests on the size of training set. This project tries to speed up the overall process. It presents various methods to generate multilingual versions of a first functionnal system, using statistical machine translation. Information from the source data is projected to another language in order to create similar target data, which then reduces the upcoming development time. Many approaches were tested and analysed. In particular, a method that regroups data in clusters before reordering the associated translation candidates shows promising results.

No description available.

Many approaches to solving tasks in the field of Natural Language Processing (NLP) use syntactic dependency trees (SDTs) as a feature to represent the latent nonlinear structure within sentences. Recently, work in parsing sentences to graph-based structures which encode semantic relationships between words—called semantic dependency graphs (SDGs)—has gained interest. This thesis seeks to explore the use of SDGs in place of and alongside SDTs within a relation classification system based on long short-term memory (LSTM) neural networks. Two methods for handling the information in these graphs are presented and compared between two SDG formalisms. Three new relation extraction system architectures have been created based on these methods and are compared to a recent state-of-the-art LSTM-based system, showing comparable results when semantic dependencies are used to enhance syntactic dependencies, but with significantly fewer training parameters.