361 |
Cluster Analysis with Meaning : Detecting Texts that Convey the Same Message / Klusteranalys med mening : Detektering av texter som uttrycker samma sakÖhrström, Fredrik January 2018 (has links)
Textual duplicates can be hard to detect as they differ in words but have similar semantic meaning. At Etteplan, a technical documentation company, they have many writers that accidentally re-write existing instructions explaining procedures. These "duplicates" clutter the database. This is not desired because it is duplicate work. The condition of the database will only deteriorate as the company expands. This thesis attempts to map where the problem is worst, and also how to calculate how many duplicates there are. The corpus is small, but written in a controlled natural language called Simplified Technical English. The method uses document embeddings from doc2vec and clustering by use of HDBSCAN* and validation using Density-Based Clustering Validation index (DBCV), to chart the problems. A survey was sent out to try to determine a threshold value of when documents stop being duplicates, and then using this value, a theoretical duplicate count was calculated.
|
362 |
Surface Realization Using a Featurized Syntactic Statistical Language ModelPacker, Thomas L. 13 March 2006 (has links)
An important challenge in natural language surface realization is the generation of grammatical sentences from incomplete sentence plans. Realization can be broken into a two-stage process consisting of an over-generating rule-based module followed by a ranker that outputs the most probable candidate sentence based on a statistical language model. Thus far, an n-gram language model has been evaluated in this context. More sophisticated syntactic knowledge is expected to improve such a ranker. In this thesis, a new language model based on featurized functional dependency syntax was developed and evaluated. Generation accuracies and cross-entropy for the new language model did not beat the comparison bigram language model.
|
363 |
Élaboration d'ontologies médicales pour une approche multi-agents d'aide à la décision clinique / A multi-agent framework for the development of medical ontologies in clinical decision makingShen, Ying 20 March 2015 (has links)
La combinaison du traitement sémantique des connaissances (Semantic Processing of Knowledge) et de la modélisation des étapes de raisonnement (Modeling Steps of Reasoning), utilisés dans le domaine clinique, offrent des possibilités intéressantes, nécessaires aussi, pour l’élaboration des ontologies médicales, utiles à l'exercice de cette profession. Dans ce cadre, l'interrogation de banques de données médicales multiples, comme MEDLINE, PubMed… constitue un outil précieux mais insuffisant car elle ne permet pas d'acquérir des connaissances facilement utilisables lors d’une démarche clinique. En effet, l'abondance de citations inappropriées constitue du bruit et requiert un tri fastidieux, incompatible avec une pratique efficace de la médecine.Dans un processus itératif, l'objectif est de construire, de façon aussi automatisée possible, des bases de connaissances médicales réutilisables, fondées sur des ontologies et, dans cette thèse, nous développons une série d'outils d'acquisition de connaissances qui combinent des opérateurs d'analyse linguistique et de modélisation de la clinique, fondés sur une typologie des connaissances mises en œuvre, et sur une implémentation des différents modes de raisonnement employés. La connaissance ne se résume pas à des informations issues de bases de données ; elle s’organise grâce à des opérateurs cognitifs de raisonnement qui permettent de la rendre opérationnelle dans le contexte intéressant le praticien.Un système multi-agents d’aide à la décision clinique (SMAAD) permettra la coopération et l'intégration des différents modules entrant dans l'élaboration d'une ontologie médicale et les sources de données sont les banques médicales, comme MEDLINE, et des citations extraites par PubMed ; les concepts et le vocabulaire proviennent de l'Unified Medical Language System (UMLS).Concernant le champ des bases de connaissances produites, la recherche concerne l'ensemble de la démarche clinique : le diagnostic, le pronostic, le traitement, le suivi thérapeutique de différentes pathologies, dans un domaine médical donné.Différentes approches et travaux sont recensés, dans l’état de question, et divers paradigmes sont explorés : 1) l'Evidence Base Medicine (une médecine fondée sur des indices). Un indice peut se définir comme un signe lié à son mode de mise en œuvre ; 2) Le raisonnement à partir de cas (RàPC) se fonde sur l'analogie de situations cliniques déjà rencontrées ; 3) Différentes approches sémantiques permettent d'implémenter les ontologies.Sur l’ensemble, nous avons travaillé les aspects logiques liés aux opérateurs cognitifs de raisonnement utilisés et nous avons organisé la coopération et l'intégration des connaissances exploitées durant les différentes étapes du processus clinique (diagnostic, pronostic, traitement, suivi thérapeutique). Cette intégration s’appuie sur un SMAAD : système multi-agent d'aide à la décision. / The combination of semantic processing of knowledge and modelling steps of reasoning employed in the clinical field offers exciting and necessary opportunities to develop ontologies relevant to the practice of medicine. In this context, multiple medical databases such as MEDLINE, PubMed are valuable tools but not sufficient because they cannot acquire the usable knowledge easily in a clinical approach. Indeed, abundance of inappropriate quotations constitutes the noise and requires a tedious sort incompatible with the practice of medicine.In an iterative process, the objective is to build an approach as automated as possible, the reusable medical knowledge bases is founded on an ontology of the concerned fields. In this thesis, the author will develop a series of tools for knowledge acquisition combining the linguistic analysis operators and clinical modelling based on the implemented knowledge typology and an implementation of different forms of employed reasoning. Knowledge is not limited to the information from data, but also and especially on the cognitive operators of reasoning for making them operational in the context relevant to the practitioner.A multi-agent system enables the integration and cooperation of the various modules used in the development of a medical ontology.The data sources are from medical databases such as MEDLINE, the citations retrieved by PubMed, and the concepts and vocabulary from the Unified Medical Language System (UMLS).Regarding the scope of produced knowledge bases, the research concerns the entire clinical process: diagnosis, prognosis, treatment, and therapeutic monitoring of various diseases in a given medical field.It is essential to identify the different approaches and the works already done.Different paradigms will be explored: 1) Evidence Based Medicine. An index can be defined as a sign related to its mode of implementation; 2) Case-based reasoning, which based on the analogy of clinical situations already encountered; 3) The different semantic approaches which are used to implement ontologies.On the whole, we worked on logical aspects related to cognitive operators of used reasoning, and we organized the cooperation and integration of exploited knowledge during the various stages of the clinical process (diagnosis, prognosis, treatment, therapeutic monitoring). This integration is based on a SMAAD: multi-agent system for decision support.
|
364 |
Multi-Agent User-Centric Specialization and Collaboration for Information RetrievalMooman, Abdelniser January 2012 (has links)
The amount of information on the World Wide Web (WWW) is rapidly growing in pace and topic diversity. This has made it increasingly difficult, and often frustrating, for information seekers to retrieve the content they are looking for as information retrieval systems (e.g., search engines) are unable to decipher the relevance of the retrieved information as it pertains to the information they are searching for.
This issue can be decomposed into two aspects: 1) variability of information relevance as it pertains to an information seeker. In other words, different information seekers may enter the same search text, or keywords, but expect completely different results. It is therefore, imperative that information retrieval systems possess an ability to incorporate a model of the information seeker in order to estimate the relevance and context of use of information before presenting results. Of course, in this context, by a model we mean the capture of trends in the information seeker's search behaviour. This is what many researchers refer to as the personalized search. 2) Information diversity. Information available on the World Wide Web today spans multitudes of inherently overlapping topics, and it is difficult for any information retrieval system to decide effectively on the relevance of the information retrieved in response to an information seeker's query. For example, the information seeker who wishes to use WWW to learn about a cure for a certain illness would receive a more relevant answer if the search engine was optimized into such domains of topics. This is what is being referred to in the WWW nomenclature as a 'specialized search'.
This thesis maintains that the information seeker's search is not intended to be completely random and therefore tends to portray itself as consistent patterns of behaviour. Nonetheless, this behaviour, despite being consistent, can be quite complex to capture. To accomplish this goal the thesis proposes a Multi-Agent Personalized Information Retrieval with Specialization Ontology (MAPIRSO). MAPIRSO offers a complete learning framework that is able to model the end user's search behaviour and interests and to organize information into categorized domains so as to ensure maximum relevance of its responses as they pertain to the end user queries. Specialization and personalization are accomplished using a group of collaborative agents. Each agent employs a Reinforcement Learning (RL) strategy to capture end user's behaviour and interests. Reinforcement learning allows the agents to evolve their knowledge of the end user behaviour and interests as they function to serve him or her. Furthermore, REL allows each agent to adapt to changes in an end user's behaviour and interests.
Specialization is the process by which new information domains are created based on existing information topics, allowing new kinds of content to be built exclusively for information seekers. One of the key characteristics of specialization domains is the seeker centric - which allows intelligent agents to create new information based on the information seekers' feedback and their behaviours.
Specialized domains are created by intelligent agents that collect information from a specific domain topic. The task of these specialized agents is to map the user's query to a repository of specific domains in order to present users with relevant information. As a result, mapping users' queries to only relevant information is one of the fundamental challenges in Artificial Intelligent (AI) and machine learning research.
Our approach employs intelligent cooperative agents that specialize in building personalized ontology information domains that pertain to each information seeker's specific needs. Specializing and categorizing information into unique domains is one of the challenge areas that have been addressed and various proposed solutions were evaluated and adopted to address growing information. However, categorizing information into unique domains does not satisfy each individualized information seeker. Information seekers might search for similar topics, but each would have different interests. For example, medical information of a specific medical domain has different importance to both the doctor and patients. The thesis presents a novel solution that will resolve the growing and diverse information by building seeker centric specialized information domains that are personalized through the information seekers' feedback and behaviours. To address this challenge, the research examines the fundamental components that constitute the specialized agent: an intelligent machine learning system, user input queries, an intelligent agent, and information resources constructed through specialized domains.
Experimental work is reported to demonstrate the efficiency of the proposed solution in addressing the overlapping information growth. The experimental work utilizes extensive user-centric specialized domain topics. This work employs personalized and collaborative multi learning agents and ontology techniques thereby enriching the queries and domains of the user.
Therefore, experiments and results have shown that building specialized ontology domains, pertinent to the information seekers' needs, are more precise and efficient compared to other information retrieval applications and existing search engines.
|
365 |
Multi-Agent User-Centric Specialization and Collaboration for Information RetrievalMooman, Abdelniser January 2012 (has links)
The amount of information on the World Wide Web (WWW) is rapidly growing in pace and topic diversity. This has made it increasingly difficult, and often frustrating, for information seekers to retrieve the content they are looking for as information retrieval systems (e.g., search engines) are unable to decipher the relevance of the retrieved information as it pertains to the information they are searching for.
This issue can be decomposed into two aspects: 1) variability of information relevance as it pertains to an information seeker. In other words, different information seekers may enter the same search text, or keywords, but expect completely different results. It is therefore, imperative that information retrieval systems possess an ability to incorporate a model of the information seeker in order to estimate the relevance and context of use of information before presenting results. Of course, in this context, by a model we mean the capture of trends in the information seeker's search behaviour. This is what many researchers refer to as the personalized search. 2) Information diversity. Information available on the World Wide Web today spans multitudes of inherently overlapping topics, and it is difficult for any information retrieval system to decide effectively on the relevance of the information retrieved in response to an information seeker's query. For example, the information seeker who wishes to use WWW to learn about a cure for a certain illness would receive a more relevant answer if the search engine was optimized into such domains of topics. This is what is being referred to in the WWW nomenclature as a 'specialized search'.
This thesis maintains that the information seeker's search is not intended to be completely random and therefore tends to portray itself as consistent patterns of behaviour. Nonetheless, this behaviour, despite being consistent, can be quite complex to capture. To accomplish this goal the thesis proposes a Multi-Agent Personalized Information Retrieval with Specialization Ontology (MAPIRSO). MAPIRSO offers a complete learning framework that is able to model the end user's search behaviour and interests and to organize information into categorized domains so as to ensure maximum relevance of its responses as they pertain to the end user queries. Specialization and personalization are accomplished using a group of collaborative agents. Each agent employs a Reinforcement Learning (RL) strategy to capture end user's behaviour and interests. Reinforcement learning allows the agents to evolve their knowledge of the end user behaviour and interests as they function to serve him or her. Furthermore, REL allows each agent to adapt to changes in an end user's behaviour and interests.
Specialization is the process by which new information domains are created based on existing information topics, allowing new kinds of content to be built exclusively for information seekers. One of the key characteristics of specialization domains is the seeker centric - which allows intelligent agents to create new information based on the information seekers' feedback and their behaviours.
Specialized domains are created by intelligent agents that collect information from a specific domain topic. The task of these specialized agents is to map the user's query to a repository of specific domains in order to present users with relevant information. As a result, mapping users' queries to only relevant information is one of the fundamental challenges in Artificial Intelligent (AI) and machine learning research.
Our approach employs intelligent cooperative agents that specialize in building personalized ontology information domains that pertain to each information seeker's specific needs. Specializing and categorizing information into unique domains is one of the challenge areas that have been addressed and various proposed solutions were evaluated and adopted to address growing information. However, categorizing information into unique domains does not satisfy each individualized information seeker. Information seekers might search for similar topics, but each would have different interests. For example, medical information of a specific medical domain has different importance to both the doctor and patients. The thesis presents a novel solution that will resolve the growing and diverse information by building seeker centric specialized information domains that are personalized through the information seekers' feedback and behaviours. To address this challenge, the research examines the fundamental components that constitute the specialized agent: an intelligent machine learning system, user input queries, an intelligent agent, and information resources constructed through specialized domains.
Experimental work is reported to demonstrate the efficiency of the proposed solution in addressing the overlapping information growth. The experimental work utilizes extensive user-centric specialized domain topics. This work employs personalized and collaborative multi learning agents and ontology techniques thereby enriching the queries and domains of the user.
Therefore, experiments and results have shown that building specialized ontology domains, pertinent to the information seekers' needs, are more precise and efficient compared to other information retrieval applications and existing search engines.
|
366 |
Relation Classification using Semantically-Enhanced Syntactic Dependency Paths : Combining Semantic and Syntactic Dependencies for Relation Classification using Long Short-Term Memory NetworksCapshaw, Riley January 2018 (has links)
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.
|
367 |
Atribuição automática de autoria de obras da literatura brasileira / Atribuição automática de autoria de obras da literatura brasileiraNobre Neto, Francisco Dantas 19 January 2010 (has links)
Made available in DSpace on 2015-05-14T12:36:48Z (GMT). No. of bitstreams: 1
arquivototal.pdf: 1280792 bytes, checksum: d335d67b212e054f48f0e8bca0798fe5 (MD5)
Previous issue date: 2010-01-19 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Authorship attribution consists in categorizing an unknown document among
some classes of authors previously selected. Knowledge about authorship of a
text can be useful when it is required to detect plagiarism in any literary
document or to properly give the credits to the author of a book. The most
intuitive form of human analysis of a text is by selecting some characteristics
that it has. The study of selecting attributes in any written document, such as
average word length and vocabulary richness, is known as stylometry. For
human analysis of an unknown text, the authorship discovery can take months,
also becoming tiring activity. Some computational tools have the functionality of
extracting such characteristics from the text, leaving the subjective analysis to
the researcher. However, there are computational methods that, in addition to
extract attributes, make the authorship attribution, based in the characteristics
gathered in the text. Techniques such as neural network, decision tree and
classification methods have been applied to this context and presented results
that make them relevant to this question. This work presents a data
compression method, Prediction by Partial Matching (PPM), as a solution of the
authorship attribution problem of Brazilian literary works. The writers and works
selected to compose the authors database were, mainly, by their representative
in national literature. Besides, the availability of the books has also been
considered. The PPM performs the authorship identification without any
subjective interference in the text analysis. This method, also, does not make
use of attributes presents in the text, differently of others methods. The correct
classification rate obtained with PPM, in this work, was approximately 93%,
while related works exposes a correct rate between 72% and 89%. In this work,
was done, also, authorship attribution with SVM approach. For that, were
selected attributes in the text divided in two groups, one word based and other in
function-words frequency, obtaining a correct rate of 36,6% and 88,4%,
respectively. / Atribuição de autoria consiste em categorizar um documento desconhecido
dentre algumas classes de autores previamente selecionadas. Saber a autoria
de um texto pode ser útil quando é necessário detectar plágio em alguma obra
literária ou dar os devidos créditos ao autor de um livro. A forma mais intuitiva
ao ser humano para se analisar um texto é selecionando algumas
características que ele possui. O estudo de selecionar atributos em um
documento escrito, como tamanho médio das palavras e riqueza vocabular, é
conhecido como estilometria. Para análise humana de um texto desconhecido,
descobrir a autoria pode demandar meses, além de se tornar uma tarefa
cansativa. Algumas ferramentas computacionais têm a funcionalidade de extrair
tais características do texto, deixando a análise subjetiva para o pesquisador.
No entanto, existem métodos computacionais que, além de extrair atributos,
atribuem a autoria baseado nas características colhidas ao longo do texto.
Técnicas como redes neurais, árvores de decisão e métodos de classificação já
foram aplicados neste contexto e apresentaram resultados que os tornam
relevantes para tal questão. Este trabalho apresenta um método de compressão
de dados, o Prediction by Partial Matching (PPM), para solução do problema de
atribuição de autoria de obras da literatura brasileira. Os escritores e obras
selecionados para compor o banco de autores se deram, principalmente, pela
representatividade que possuem na literatura nacional. Além disso, a
disponibilidade dos livros em formato eletrônico também foi considerada. O
PPM realiza a identificação de autoria sem ter qualquer interferência subjetiva
na análise do texto. Este método, também, não faz uso de atributos presentes
ao longo do texto, diferentemente de outros métodos. A taxa de classificação
correta alcançada com o PPM, neste trabalho, foi de aproximadamente 93%,
enquanto que trabalhos relacionados mostram uma taxa de acerto entre 72% e
89%. Neste trabalho, também foi realizado atribuição de autoria com a
abordagem SVM. Para isso, foram selecionados atributos no texto dividido em
dois tipos, sendo um baseado em palavras e o outro na contagem de palavrasfunção,
obtendo uma taxa de acerto de 36,6% e 88,4%, respectivamente.
|
368 |
Elaboration de ressources électroniques pour les noms composés de type N (E+DET=G) N=G du grec moderne / The N (E + DET=G) N=G compound nouns in Modern GreekKyriakopoulou, Anthoula 25 March 2011 (has links)
L'objectif de cette recherche est la construction manuelle de ressources lexicales pour les noms composés grecs qui sont définis par la structure morphosyntaxique : Nom (E+Déterminant au génitif) Nom au génitif, notés N (E+DET:G) N:G (e.g. ζώνη ασφαλείας/ceinture de sécurité). Les ressources élaborées peuvent être utilisées pour leur reconnaissance lexicale automatique dans les textes écrits et dans d'autres applications du TAL. Notre travail s'inscrit dans la perspective de l'élaboration du lexique-grammaire général du grec moderne en vue de l'analyse automatique des textes écrits. Le cadre théorique et méthodologique de cette étude est celui du lexique-grammaire (M. Gross 1975, 1977), qui s'appuie sur la grammaire transformationnelle harisienne.Notre travail s'organise en cinq parties. Dans la première partie, nous délimitons l'objet de notre travail tout en essayant de définir la notion fondamentale qui régit notre étude, à savoir celle de figement. Dans la deuxième partie, nous présentons la méthodologie utilisée pour le recensement de nos données lexicales et nous étudions les phénomènes de variation observés au sein des noms composés de type N (E+DET:G) N:G. La troisième partie est consacrée à la présentation des différentes sous-catégories des N (E+DET:G) N:G identifiées lors de l'étape du recensement et à l'étude de leur structure lexicale interne. La quatrième partie porte sur l'étude syntaxico-sémantique des N (E+DET:G) N:G. Enfin, dans la cinquième partie, nous présentons les différentes méthodes de représentation formalisée que nous proposons pour nos données lexicales en vue de leur reconnaissance lexicale automatique dans les textes écrits. Des échantillons représentatifs des ressources élaborées sont présentés en Annexe / The object of this research is the manual construction of lexical resources for the Greek compound nouns defined by the following morphosyntactic structure : Noun (E+Determiner in genitive) Noun in genitive, (N (E+DET:G) N:G) (e.g. ζώνη ασφαλείας/safety belt). The elaborated resources may be used for their automatic recognition in written texts and other NLP applications. Our study is part of the general lexicon-grammar for Modern Greek in view of automatic processing of written texts. Our theoretical and methodological framework is that of lexicon-grammar (M. Gross 1975, 1977), based on the Transformational Grammar principles defined by Z. S. Harris. Our study is organised into five parts. In the first part, we give an overview of the core notion governing our research : the notions of (fixed) multiword expression (MWE). In the second part, we present the methodology used to collect our lexical data and we study the variation phenomena observed within the framework of the N (E+DET:G) N:G. The third part is dedicated to the presentation of the different N (E+DET:G) N:G categories identified in the listing phase qnd to the study of their lexical composition. The fourth concerns the syntactical and semantic study of the N (E+DET:G) N:G. Finally, the fifth part deals with the formal representation methods we propose for our lexical data in view of their lexical recognition in Greek written texts. Representative samples of the elaborated resources are illustrated in Appendix
|
369 |
Word2vec modely s přidanou kontextovou informací / Word2vec Models with Added Context InformationŠůstek, Martin January 2017 (has links)
This thesis is concerned with the explanation of the word2vec models. Even though word2vec was introduced recently (2013), many researchers have already tried to extend, understand or at least use the model because it provides surprisingly rich semantic information. This information is encoded in N-dim vector representation and can be recall by performing some operations over the algebra. As an addition, I suggest a model modifications in order to obtain different word representation. To achieve that, I use public picture datasets. This thesis also includes parts dedicated to word2vec extension based on convolution neural network.
|
370 |
Natural Language Processing using Deep Learning in Social MediaGiménez Fayos, María Teresa 02 September 2021 (has links)
[ES] En los últimos años, los modelos de aprendizaje automático profundo (AP) han revolucionado los sistemas de procesamiento de lenguaje natural (PLN).
Hemos sido testigos de un avance formidable en las capacidades de estos sistemas y actualmente podemos encontrar sistemas que integran modelos PLN de manera ubicua.
Algunos ejemplos de estos modelos con los que interaccionamos a diario incluyen modelos que determinan la intención de la persona que escribió un texto, el sentimiento que pretende comunicar un tweet o nuestra ideología política a partir de lo que compartimos en redes sociales.
En esta tesis se han propuestos distintos modelos de PNL que abordan tareas que estudian el texto que se comparte en redes sociales. En concreto, este trabajo se centra en dos tareas fundamentalmente: el análisis de sentimientos y el reconocimiento de la personalidad de la persona autora de un texto.
La tarea de analizar el sentimiento expresado en un texto es uno de los problemas principales en el PNL y consiste en determinar la polaridad que un texto pretende comunicar. Se trata por lo tanto de una tarea estudiada en profundidad de la cual disponemos de una vasta cantidad de recursos y modelos.
Por el contrario, el problema del reconocimiento de personalidad es una tarea revolucionaria que tiene como objetivo determinar la personalidad de los usuarios considerando su estilo de escritura. El estudio de esta tarea es más marginal por lo que disponemos de menos recursos para abordarla pero que no obstante presenta un gran potencial.
A pesar de que el enfoque principal de este trabajo fue el desarrollo de modelos de aprendizaje profundo, también hemos propuesto modelos basados en recursos lingüísticos y modelos clásicos del aprendizaje automático. Estos últimos modelos nos han permitido explorar las sutilezas de distintos elementos lingüísticos como por ejemplo el impacto que tienen las emociones en la clasificación correcta del sentimiento expresado en un texto.
Posteriormente, tras estos trabajos iniciales se desarrollaron modelos AP, en particular, Redes neuronales convolucionales (RNC) que fueron aplicadas a las tareas previamente citadas. En el caso del reconocimiento de la personalidad, se han comparado modelos clásicos del aprendizaje automático con modelos de aprendizaje profundo, pudiendo establecer una comparativa bajo las mismas premisas.
Cabe destacar que el PNL ha evolucionado drásticamente en los últimos años gracias al desarrollo de campañas de evaluación pública, donde múltiples equipos de investigación comparan las capacidades de los modelos que proponen en las mismas condiciones. La mayoría de los modelos presentados en esta tesis fueron o bien evaluados mediante campañas de evaluación públicas, o bien emplearon la configuración de una campaña pública previamente celebrada. Siendo conscientes, por lo tanto, de la importancia de estas campañas para el avance del PNL, desarrollamos una campaña de evaluación pública cuyo objetivo era clasificar el tema tratado en un tweet, para lo cual recogimos y etiquetamos un nuevo conjunto de datos.
A medida que avanzabamos en el desarrollo del trabajo de esta tesis, decidimos estudiar en profundidad como las RNC se aplicaban a las tareas de PNL.
En este sentido, se exploraron dos líneas de trabajo.
En primer lugar, propusimos un método de relleno semántico para RNC, que plantea una nueva manera de representar el texto para resolver tareas de PNL. Y en segundo lugar, se introdujo un marco teórico para abordar una de las críticas más frecuentes del aprendizaje profundo, el cual es la falta de interpretabilidad. Este marco busca visualizar qué patrones léxicos, si los hay, han sido aprendidos por la red para clasificar un texto. / [CA] En els últims anys, els models d'aprenentatge automàtic profund (AP) han revolucionat els sistemes de processament de llenguatge natural (PLN).
Hem estat testimonis d'un avanç formidable en les capacitats d'aquests sistemes i actualment podem trobar sistemes que integren models PLN de manera ubiqua.
Alguns exemples d'aquests models amb els quals interaccionem diàriament inclouen models que determinen la intenció de la persona que va escriure un text, el sentiment que pretén comunicar un tweet o la nostra ideologia política a partir del que compartim en xarxes socials.
En aquesta tesi s'han proposats diferents models de PNL que aborden tasques que estudien el text que es comparteix en xarxes socials. En concret, aquest treball se centra en dues tasques fonamentalment: l'anàlisi de sentiments i el reconeixement de la personalitat de la persona autora d'un text.
La tasca d'analitzar el sentiment expressat en un text és un dels problemes principals en el PNL i consisteix a determinar la polaritat que un text pretén comunicar. Es tracta per tant d'una tasca estudiada en profunditat de la qual disposem d'una vasta quantitat de recursos i models.
Per contra, el problema del reconeixement de la personalitat és una tasca revolucionària que té com a objectiu determinar la personalitat dels usuaris considerant el seu estil d'escriptura. L'estudi d'aquesta tasca és més marginal i en conseqüència disposem de menys recursos per abordar-la però no obstant i això presenta un gran potencial.
Tot i que el fouc principal d'aquest treball va ser el desenvolupament de models d'aprenentatge profund, també hem proposat models basats en recursos lingüístics i models clàssics de l'aprenentatge automàtic. Aquests últims models ens han permès explorar les subtileses de diferents elements lingüístics com ara l'impacte que tenen les emocions en la classificació correcta del sentiment expressat en un text.
Posteriorment, després d'aquests treballs inicials es van desenvolupar models AP, en particular, Xarxes neuronals convolucionals (XNC) que van ser aplicades a les tasques prèviament esmentades. En el cas de el reconeixement de la personalitat, s'han comparat models clàssics de l'aprenentatge automàtic amb models d'aprenentatge profund la qual cosa a permet establir una comparativa de les dos aproximacions sota les mateixes premisses.
Cal remarcar que el PNL ha evolucionat dràsticament en els últims anys gràcies a el desenvolupament de campanyes d'avaluació pública on múltiples equips d'investigació comparen les capacitats dels models que proposen sota les mateixes condicions. La majoria dels models presentats en aquesta tesi van ser o bé avaluats mitjançant campanyes d'avaluació públiques, o bé s'ha emprat la configuració d'una campanya pública prèviament celebrada. Sent conscients, per tant, de la importància d'aquestes campanyes per a l'avanç del PNL, vam desenvolupar una campanya d'avaluació pública on l'objectiu era classificar el tema tractat en un tweet, per a la qual cosa vam recollir i etiquetar un nou conjunt de dades.
A mesura que avançàvem en el desenvolupament del treball d'aquesta tesi, vam decidir estudiar en profunditat com les XNC s'apliquen a les tasques de PNL. En aquest sentit, es van explorar dues línies de treball.En primer lloc, vam proposar un mètode d'emplenament semàntic per RNC, que planteja una nova manera de representar el text per resoldre tasques de PNL. I en segon lloc, es va introduir un marc teòric per abordar una de les crítiques més freqüents de l'aprenentatge profund, el qual és la falta de interpretabilitat. Aquest marc cerca visualitzar quins patrons lèxics, si n'hi han, han estat apresos per la xarxa per classificar un text. / [EN] In the last years, Deep Learning (DL) has revolutionised the potential of automatic systems that handle Natural Language Processing (NLP) tasks.
We have witnessed a tremendous advance in the performance of these systems. Nowadays, we found embedded systems ubiquitously, determining the intent of the text we write, the sentiment of our tweets or our political views, for citing some examples.
In this thesis, we proposed several NLP models for addressing tasks that deal with social media text. Concretely, this work is focused mainly on Sentiment Analysis and Personality Recognition tasks.
Sentiment Analysis is one of the leading problems in NLP, consists of determining the polarity of a text, and it is a well-known task where the number of resources and models proposed is vast.
In contrast, Personality Recognition is a breakthrough task that aims to determine the users' personality using their writing style, but it is more a niche task with fewer resources designed ad-hoc but with great potential.
Despite the fact that the principal focus of this work was on the development of Deep Learning models, we have also proposed models based on linguistic resources and classical Machine Learning models. Moreover, in this more straightforward setup, we have explored the nuances of different language devices, such as the impact of emotions in the correct classification of the sentiment expressed in a text.
Afterwards, DL models were developed, particularly Convolutional Neural Networks (CNNs), to address previously described tasks. In the case of Personality Recognition, we explored the two approaches, which allowed us to compare the models under the same circumstances.
Noteworthy, NLP has evolved dramatically in the last years through the development of public evaluation campaigns, where multiple research teams compare the performance of their approaches under the same conditions. Most of the models here presented were either assessed in an evaluation task or either used their setup. Recognising the importance of this effort, we curated and developed an evaluation campaign for classifying political tweets.
In addition, as we advanced in the development of this work, we decided to study in-depth CNNs applied to NLP tasks.
Two lines of work were explored in this regard.
Firstly, we proposed a semantic-based padding method for CNNs, which addresses how to represent text more appropriately for solving NLP tasks. Secondly, a theoretical framework was introduced for tackling one of the most frequent critics of Deep Learning: interpretability. This framework seeks to visualise what lexical patterns, if any, the CNN is learning in order to classify a sentence.
In summary, the main achievements presented in this thesis are:
- The organisation of an evaluation campaign for Topic Classification from texts gathered from social media.
- The proposal of several Machine Learning models tackling the Sentiment Analysis task from social media. Besides, a study of the impact of linguistic devices such as figurative language in the task is presented.
- The development of a model for inferring the personality of a developer provided the source code that they have written.
- The study of Personality Recognition tasks from social media following two different approaches, models based on machine learning algorithms and handcrafted features, and models based on CNNs were proposed and compared both approaches.
- The introduction of new semantic-based paddings for optimising how the text was represented in CNNs.
- The definition of a theoretical framework to provide interpretable information to what CNNs were learning internally. / Giménez Fayos, MT. (2021). Natural Language Processing using Deep Learning in Social Media [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/172164 / TESIS
|
Page generated in 0.7009 seconds