Catálogo de publicaciones - libros

We use an evolutionary process to seek a specialized powerful rule of Cellular Automata (CA) among a set of best rules for extracting edges in a given black-white image. This best set of local rules determines the future state of CA in an asynchronous way. The Genetic Algorithm (GA) is applied to search the best CA rules that can realize better the edge detection.

Humour is one of the most interesting and puzzling aspects of human behaviour, and it has been rightfully argued that it plays an important role in an individual’s development, as well as in interpersonal communication. Research on this topic has received a significant amount of attention from fields as diverse as linguistics, philosophy, psychology and sociology, and recent years have also seen attempts to build computational models for humour generation and recognition.

We consider the problem of automatic classification of text documents, in particular, scientific abstracts and use two types of classifiers: ordinal and numerical. For the first type we use a fuzzy extension of the Borda voting method while for the second type we use a fuzzy Borda method in combination with the semantic grading.

Commonly used vocabulary in Indian language documents found on the web contain a number of words that have Sanskrit, Persian or English origin. However, such words may be written in different scripts with slight variations in spelling and morphology. In this paper we explore approximate string matching techniques to exploit this situation of relatively large number of cognates among Indian languages, which are higher when compared to an Indian language and a non-Indian language. We present an approach to identify cognates and make use of them for improving dictionary based CLIR when the query and documents both belong to two different Indian languages. We conduct experiments using a Hindi document collection and a set of Telugu queries and report the improvement due to cognate recognition and translation.

Nowadays, the multimodal nature of the World Wide Web is an evidence. Web sites which include video files, pictures, music and text have become widespread. Furthermore, multimodal collections in several languages demand to apply multilingual information retrieval strategies. This paper describes a new retrieval technique applied on a multimodal and multilingual system that have been tested on two different multilingual image collections. The system applies several machine translators and implements some novel heuristics. These heuristics explore a variety of ways to combine the translations obtained from the given set of translators, and the configuration of the retrieval model by using different weighting functions, and also studying the effect of pseudo-relevance feedback (PRF) on this domain. Our results show interesting effects by these variations, allowing the determination of the parameters for the best retrieval model on this data and reporting the loss in performance on each language.

The paper reports on recent experiments in cross-lingual document processing (with a case study for Bulgarian-English-Romanian language pairs) and brings evidence on the benefits of using linguistic ontologies for achieving, with a high level of accuracy, difficult tasks in NLP such as word alignment, word sense disambiguation, document classification, cross-language information retrieval, etc. We provide brief descriptions of the parallel corpus we used, the multilingual lexical ontology which supports our research, the word alignment and word sense disambiguation systems we developed and a preliminary report on an ongoing development of a system for cross-lingual text-classification which takes advantage of these multilingual technologies. Unlike the keyword-based methods in document processing, the concept-based methods are supposed to better exploit the semantic information contained in a particular document and thus to provide more accurate results.

In this paper we introduce the NTCIR6 Opinion Analysis Pilot Task, information about the Chinese, Japanese, and English data, plans for future opinion analysis tasks at NTCIR, and a brief overview of the evaluation results. This pilot task is a sentence-level opinion identification and polarity detection task run over data from a comparable corpus in three languages: Chinese, English, and Japanese. We have manually annotated documents for this task in each language, producing what we believe to be the first multilingual opinion analysis data set over comparable data. Six participants submitted Chinese system results, three Japanese, and six English for this pilot task. We plan to release the data to the research community, and hope to spur further research into cross-lingual opinion analysis and its use in other NLP tasks. In particular, we look forward to researchers using this data to investigate cross-cultural perspective differences based on automatic sentiment analysis.

In this paper we present some results obtained in humour classification over a corpus of Italian quotations manually extracted and tagged from the Wikiquote project. The experiments were carried out using both a multinomial Naïve Bayes classifier and a Support Vector Machine (SVM). The considered features range from single words to n -grams and sentence length. The obtained results show that it is possible to identify the funny quotes even with the simplest features (bag of words); the bayesian classifier performed better than the SVM. However, the size of the corpus size is too small to support definitive assertions.

We present a machine learning approach for classifying sentences as one-liner jokes or normal sentences. We use no deep analysis of the meaning to try to see if it is humorous, instead we rely on a combination of simple features to see if these are enough to detect humor. Features such as word overlap with other jokes, presence of words common in jokes, ambiguity and word overlap with common idioms turn out to be useful. When training and testing on equal amounts of jokes and sentences from the British National Corpus, a classification accuracy of 85% is achieved.

This paper presents a computer implementation that utilizes cross-reference ambiguity in utterances for simple conversational jokes. The approach is based on the SSTH. Using a simple script representation, it is shown that cross-reference ambiguities always satisfy the SSTH requirement for script overlap. To determine whether script opposition is present, we introduce a method that compares the concepts involved based on their semantic properties. When a given cross-reference ambiguity results in script opposition it is possible to generate a punchline based on this ambiguity. As a result of the low performance of the anaphora resolution algorithm and the data sparseness in ConceptNet the application performs moderately, but it does provide future prospects in generating conversational humour.