Search

This paper summarizes the thesis: ”Author Profiling in Social Media with Multimodal Information.” Our solution uses a multimodal approach to extracting information from written messages and images shared by users. Previous work has shown the existence of useful information for this task in these modalities; however, our proposal goes further, demonstrating the complementarity of the modalities when merging these two sources of information. To do this, we propose to transform images to texts, and with them, to have the same framework of representation for both kinds of information, which allow to achieve their fusion. Our work explores different methods for extracting information either from the text and the images. To represent the extracted information, different distributional term representations approaches were explored in order to identify the topics addressed by the user. For this purpose, an evaluation framework was proposed in order to identify the most appropriate method for this task. The results show that the textual descriptions of the images contain useful information for the author profiling task, and that the fusion of textual information with information extracted from the images increases the accuracy of this task.

Recent works have shown that it is possible to use information extracted from images to address the task of automatic gender identification. These proposals have validated their solutions using monolingual datasets, i.e., collections where images are shared by users having the same mother tongue. This paper aims to test the usefulness of images collected from users who do not share the same language. In principle, these users present cultural differences, which may be reflected in the images they share. However, a cross-cultural image-based approach would be very useful for languages where data is not available or scarce. The experiments presented demonstrate that characteristics obtained from the images, regardless of the users’ mother tongue, can be used for gender prediction. They mainly confirm the usefulness of a cross-cultural image-based approach, showing that culturally different individuals with equivalent profiles traits tend to share similar images.

The goal of Author Profiling (AP) is to identify demographic aspects (e.g., age, gender) from a given set of authors by analyzing their written texts. Recently, the AP task has gained interest in many problems related to computer forensics, psychology, marketing, but specially in those related with social media exploitation. As known, social media data is shared through a wide range of modalities (e.g., text, images and audio), representing valuable information to be exploited for extracting valuable insights from users. Nevertheless, most of the current work in AP using social media data has been devoted to analyze textual information only, and there are very few works that have started exploring the gender identification using visual information. Contrastingly, this paper focuses in exploiting the visual modality to perform both age and gender identification in social media, specifically in Twitter. Our goal is to evaluate the pertinence of using visual information in solving the AP task. Accordingly, we have extended the Twitter corpus from PAN 2014, incorporating posted images from all the users, making a distinction between tweeted and retweeted images. Performed experiments provide interesting evidence on the usefulness of visual information in comparison with traditional textual representations for the AP task.

The Author Profiling (AP) task aims to reveal as much as possible information from a given author's document (e.g., age, gender, etc.). AP is crucial for several applications, ranging from customized advertising to computer forensics, psychology, and entertainment. Nonetheless, the AP task is far from being solved, particularly in social media domains, where the nature of documents hinder the applicability of state-of-the-art text mining tools (e.g., because of spelling-grammar errors, huge vocabularies, and the presence of many out-of-vocabulary terms). Currently, most of the work in AP for social media has been devoted to the development of descriptive features, which are used under standard representations, such as the Bag-of-Words (BoW). Nevertheless, BoW-like representations have some well known shortcomings, namely: (i) the sparsity and high dimensionality of the representation, and (ii) the failure to capture relationships, other than mere occurrence, among terms. This paper focuses on the study of alternative document representations that can deal with such issues. We propose a representation for documents that capture discriminative and subprofile-specific information of terms. Under the proposed representation, terms are represented in a vector space that captures discriminative information. Then, term representations are aggregated to represent the content of a document. In this manner, documents are represented in a low-dimensional (and discriminative) space which is non-sparse. We evaluate the effectiveness of the proposed representation on several corpora from the social media domain. The proposed representation is compared to the standard BoW representation and a wide variety of state-of-the-art AP approaches. Experimental results reveal that the proposed representation outperforms most of the reference methodologies. Furthermore, we show that the proposed representation is in agreement with previous studies on handcrafted attributes for AP.

The vast amount of electronic documents available on the Internet demands for automatic tools that help people finding, organizing and easily accessing to all this information. Although current text classification methods have alleviated some of the above problems, such strategies depend on having a large and reliable set of labeled data. In order to overcome such limitation, this work proposes an alternative approach for semi-supervised text classification, which is based on a new strategy for diminishing the sensitivity to the noise contained on labeled data by means of automatic text summarization. Experimental results showed that our proposed approach outperforms traditional semi-supervised text classification techniques; additionally, our results also indicate that our approach is suitable for learning from only one labeled example per category.

During the last decades the Web has become the greatest repository of digital information. In order to organize all this information, several text categorization methods have been developed, achieving accurate results in most cases and in very different domains. Due to the recent usage of Internet as communication media, short texts such as news, tweets, blogs, and product reviews are more common every day. In this context, there are two main challenges; on the one hand, the length of these documents is short, and therefore, the word frequencies are not informative enough, making text categorization even more difficult than usual. On the other hand, topics are changing constantly at a fast rate, causing the lack of adequate amounts of training data. In order to deal with these two problems we consider a text classification method that is supported on the idea that similar documents may belong to the same category. Mainly, we propose a neighborhood consensus classification method that classifies documents by considering their own information as well as information about the category assigned to other similar documents from the same target collection. In particular, the short texts we used in our evaluation are news titles with an average of 8 words. Experimental results are encouraging; they indicate that leveraging information from similar documents helped to improve classification accuracy and that the proposed method is especially useful when labeled training resources are limited.

Question answering (QA) is the task of automatically answering a question posed in natural language. Currently, there exists several QA approaches, and, according to recent evaluation results, most of them are complementary. That is, different systems are relevant for different kinds of questions. Somehow, this fact indicates that a pertinent combination of various systems should allow to improve the individual results. This paper focuses on this problem, namely, the selection of the correct answer from a given set of responses corresponding to different QA systems. In particular, it proposes a supervised multi-stream approach that decides about the correctness of answers based on a set of features that describe: (i) the compatibility between question and answer types, (ii) the redundancy of answers across streams, as well as (iii) the overlap and non-overlap information between the question-answer pair and the support text. Experimental results are encouraging; evaluated over a set of 190 questions in Spanish and using answers from 17 different QA systems, our multi-stream QA approach could reach an estimated QA performance of 0.74, significantly outperforming the estimated performance from the best individual system (0.53) as well as the result from best traditional multi-stream QA approach (0.60).

This paper proposes a novel representation for Authorship Attribution (AA), based on Concise Semantic Analysis (CSA), which has been successfully used in Text Categorization (TC). Our approach for AA, called Document Author Representation (DAR), builds document vectors in a space of authors, calculating the relationship between textual features and authors. In order to evaluate our approach, we compare the proposed representation with conventional approaches and previous works using the c50 corpus. We found that DAR can be very useful in AA tasks, because it provides good performance on imbalanced data, getting comparable or better accuracy results.

Crosslingual text classification consists of exploiting labeled documents in a source language to classify documents in a different target language. In addition to the evident translation problem, this task also faces some difficulties caused by the cultural discrepancies manifested in both languages by means of different topic distributions. Such discrepancies make the classifier unreliable for the categorization task. In order to tackle this problem we propose to improve the classification performance by using information embedded in the own target dataset. The central idea of the proposed approach is that similar documents must belong to the same category. Therefore, it classifies the documents by considering not only their own content but also information about the assigned category to other similar documents from the same target dataset. Experimental results using three different languages evidence the appropriateness of the proposed approach.

Geographic Information Retrieval (GIR) is a specialized Information Retrieval (IR) branch that deals with information related to geographical locations. Traditional IR engines are perfectly able to retrieve the majority of the relevant documents for most geographical queries, but they have severe difficulties generating a pertinent ranking of the retrieved results, which leads to poor performance. A key reason for this ranking problem has been a lack of information. Therefore, previous GIR research has tried to fill this gap using robust geographical resources (i.e. a geographical ontology), while other research with the same aim has used relevant feedback techniques instead. This paper explores the use of Bag of Concepts (BoC; a representation where documents are considered as the union of the meanings of its terms) and Holographic Reduced Representation (HRR; a novel representation for textual structure) as re-ranking mechanisms for GIR. Our results reveal an improvement in mean average precision (MAP) when compared to the traditional vector space model, even if Pseudo Relevance Feedback is employed.

This year we evaluated our supervised answer validation method at both, the Spanish Answer Validation Exercise (AVE) and the Spanish Question Answering Main Task. This paper describes and analyzes our evaluation results from both tracks. In resume, the F-measure of the proposed method outperformed the baseline result of the AVE 2008 task by more than 100%, and enhanced the performance of our question answering system, showing a gain in accuracy of 22% for answering factoid questions. A detailed analysis of the results shows that the proposed non-overlap features are most discriminative than the traditional overlap ones. In particular, these novel features allowed increasing the F-measure result of our method by 26%.

Cross-language text classification (CLTC) aims to take advantage of existing training data from one language to construct a classifier for another language. In addition to the expected translation issues, CLTC is also complicated by the cultural distance between both languages, which causes that documents belonging to the same category concern very different topics. This paper proposes a re-classification method which purpose is to reduce the errors caused by this phenomenon by considering information from the own target language documents. Experimental results in a news corpus considering three pairs of languages and four categories demonstrated the appropriateness of the proposed method, which could improve the initial classification accuracy by up to 11%.

This paper focuses on the problem of ranking documents for Geographic Information Retrieval. It aims to demonstrate that by using some query-related example texts it is possible to improve the final ranking of the retrieved documents. Experimental results indicated that our approach could improve the MAP of some sets of retrieved documents using only two example texts.

Recent evaluation results from Geographic Information Retrieval (GIR) indicate that current information retrieval methods are effective to retrieve relevant documents for geographic queries, but they have severe difficulties to generate a pertinent ranking of them. Motivated by these results in this paper we present a novel re-ranking method, which employs information obtained through a relevance feedback process to perform a ranking refinement . Performed experiments show that the proposed method allows to improve the generated ranking from a traditional IR machine, as well as results from traditional re-ranking strategies such as query expansion via relevance feedback.

Some recent works have shown that the "perfect" selection of the best IR system per query could lead to a significant improvement on the retrieval performance. Motivated by this fact, in this paper we focus on the automatic selection of the best retrieval result from a given set of results lists generated by different IR systems. In particular, we propose five heuristic measures for evaluating the relative relevance of each result list, which take into account the redundancy and ranking of documents across the lists. Preliminary results in three different data sets, and considering 216 queries, are encouraging. They show that the proposed approach could slightly outperform the results from the best individual IR system in two out of three collections, but that it could significantly improve the average results of individual systems from all data sets. In addition, the achieved results indicate that our approach is a competitive alternative to traditional data fusion methods.

The bag of words representation (BoW), which is widely used in information retrieval (IR), represents documents and queries as word lists that do not express anything about context information. When we look for information, we find that not everything is explicitly stated in a document, so context information is needed to understand its content. This paper proposes the use of bag of concepts (BoC) and Holographic reduced representation (HRR) in IR. These representations go beyond BoW by incorporating context information to document representations. Both HRR and BoC are produced using a vector space methodology known as Random Indexing, and allow expressing additional knowledge from different sources. Our experiments have shown the feasibility of the representations and improved the mean average precision by up to 7% when they are compared with the traditional vector space model.

Nowadays there exist several kinds of question answering systems. According to recent evaluation results, most of these systems are complementary (i.e., each one is better than the others in answering some specific type of questions). This fact indicates that a pertinent combination of various systems may allow improving the best individual result. This paper focuses on this problem. It proposes using an answer validation method to handle this combination. The main advantage of this approach is that it does not rely on internal system's features nor depend on external answer's redundancies. Experimental results confirm the appropriateness of our proposal. They mainly show that it outperforms individual system's results as well as the precision obtained by a redundancy-based combination strategy.

Current storage and processing facilities have caused the emergence of many multimedia repositories and, consequently, they have also triggered the necessity of new approaches for information retrieval. In particular, spoken document retrieval is a very complex task since existing speech recognition systems tend to generate several transcription errors (such as word substitutions, insertions and deletions). In order to deal with these errors, this paper proposes an enriched document representation based on a phonetic codification of the automatic transcriptions. This representation aims to reduce the impact of the transcription errors by representing words with similar pronunciations through the same phonetic code. Experimental results on the CL-SR corpus from the CLEF 2007 (which includes 33 test topics and 8,104 English interviews) are encouraging; our method achieved a mean average precision of 0.0795, outperforming all except one of the evaluated systems at this forum.

As any other text categorization task, authorship attribution requires a large number of training examples. These examples, which are easily obtained for most of the tasks, are particularly difficult to obtain for this case. Based on this fact, in this paper we investigate the possibility of using Web-based text mining methods for the identification of the author of a given poem. In particular, we propose a semi-supervised method that is specially suited to work with justfew training examples in order to tackle the problem of the lack of data with the same writing style. The method considers the automatic extraction of the unlabeled examples from the Web and its iterative integration into the training data set. To the knowledge of the authors, a semi-supervised method which makes use of the Web as support lexical resource has not been previously employed in this task. The results obtained on poem categorization show that this method may improve the classification accuracy and it is appropriate to handle the attribution of short documents.

One major problem in multilingual Question Answering (QA) is the combination of answers obtained from different languages into one single ranked list. This paper proposes a new method for tackling this problem. This method is founded on a graph-based ranking approach inspired in the popular Google's PageRank algorithm. Experimental results demonstrate that the proposed method outperforms other current techniques for answer fusion, and also evidence the advantages of multilingual QA over the traditional monolingual approach.

This paper describes a QA system centered in a full data-driven architecture. It applies machine learning and text mining techniques to identify the most probable answers to factoid and definition questions respectively. Its major quality is that it mainly relies on the use of lexical information and avoids applying any complex language processing resources such as named entity classifiers, parsers and ontologies. Experimental results on the Spanish Question Answering task at CLEF 2006 show that the proposed architecture can be a practical solution for monolingual question answering by reaching a precision as high as 51%.

This work is a revised version of the paper “INAOE-UPV Joint Participation at CLEF 2005: Experiments in Monolingual Question Answering”, previously published in the CLEF 2005 working notes (www.clef-campaign.org/2005/working_notes/). This paper describes a full data-driven system for question answering. The system uses pattern matching and statistical techniques to identify the relevant passages as well as the candidate answers for factoid and definition questions. Since it does not consider any sophisticated linguistic analysis of questions and answers, it can be applied to different languages without requiring major adaptation changes. Experimental results on Spanish, Italian and French demonstrate that the proposed approach can be a convenient strategy for monolingual and multilingual question answering. CONACYT (Project Grant 43990); R2D2 (CICYTTIC2003-07158-C04-03); ICT EU-India (ALA/95/23/2003/077-054)

This paper describes a method for definition question answering based on the use of surface text patterns. The method is specially suited to answer questions about person’s positions and acronym’s descriptions. It considers two main steps. First, it applies a sequence-mining algorithm to discover a set of definition-related text patterns from the Web. Then, using these patterns, it extracts a collection of concept-description pairs from a target document database, and applies the sequence-mining algorithm to determine the most adequate answer to a given question. Experimental results on the Spanish CLEF 2005 data set indicate that this method can be a practical solution for answering this kind of definition questions, reaching a precision as high as 84%.

Passage Retrieval (PR) is typically used as the first step in current Question Answering (QA) systems. Most methods are based on the vector space model allowing the finding of relevant passages for general user needs, but failing on selecting pertinent passages for specific user questions. This paper describes a simple PR method specially suited for the QA task. This method considers the structure of the question, favoring the passages that contain the longer n-gram structures from the question. Experimental results of this method on Spanish, French and Italian show that this approach can be useful for multilingual question answering systems.

The problem of the resolution of the lexical ambiguity seems to be stuck because of the knowledge acquisition bottleneck. Therefore, it is worthwhile to investigate the possibility of using the Web as a lexical resource. This paper explores two attempts of using Web counts collected through a search engine. The first approach calculates the hits of each possible synonym of the noun to disambiguate together with the nouns of the context. In the second approach the disambiguation of a noun uses a modifier adjective as supporting evidence. A better precision than the baseline was obtained using adjective-noun pairs, even if with a low recall. A comprehensive set of weighting formulae for combining Web counts was investigated in order to give a complete picture of what are the various possibilities, and what are the formulae that work best. The comparison across different search engines was also useful: Web counts, and consequently disambiguation results, were almost identical. Moreover, the Web seems to be more effective than the WordNet Domains lexical resource if integrated rather than stand-alone.

Finding accurate information on the Web has become a challenge due to the increment in the number of documents available online. Current search engines retrieve relevant documents to general - often short - user queries, but fail extracting answers to simple factual questions in natural language. This work presents the basis of a statistical question answering system capable to find answers to factual questions in Spanish language from the Web. This approach is supported on data redundancy rather than on sophisticated linguistic analyses of either questions and candidate answers. Preliminary results show that it is feasible to find concise and accurate answers from the Web to factual questions made in Spanish language. The study also concludes that the available Spanish documents in the Web are redundant enough in order to apply statistical methods like those described in this document in order to provide better mechanisms for information access.

Nowadays there is a large amount of digital texts available for every purpose. New flexible and robust approaches are necessary for their access and analysis. This paper proposes a text exploration scheme based on hypertext, which incorporates some elements from information retrieval and text mining in order to transform the blind navigation of the hypertext into a step-by-step informed exploration. The proposed scheme is of relevance since it integrates three basic exploration functionalities, i.e. access, navigation and analysis. The paper also presents some preliminary results on the generation of hypertext from two text collections in an implementation of the scheme.

The problem of acquiring valuable information from the large amounts available today in electronic media requires automated mechanisms more natural and efficient than those already existing. The trend in the evolution of information retrieval systems goes toward systems capable of answering specific questions formulated by the user in her/his language. The expected answers from such systems are short and accurate sentences, instead of large document lists. On the other hand, the state of the art of these systems is focused -mainly- in the resolution of factual questions, whose answers are named entities (dates, quantities, proper nouns, etc). This paper proposes a model to represent source documents that are then used by question answering systems. The model is based on a representation of a document as a set of named entities (NEs) and their local lexical context. These NEs are extracted and classified automatically by an off-line process. The entities are then taken as instance concepts in an upper ontology and stored as a set of DAML+OIL resources which could be used later by question answering engines. The paper presents a case of study with a news collection in Spanish and some preliminary results.

Question Answering has become a promising research field whose aim is to provide more natural access to the information than traditional document retrieval techniques. In this work, an approach centered in the use of context at a lexical level has been followed in order to identify possible answers to short factoid questions stated by the user in natural language. The methods applied at different stages of the system as well as an architecture for question answering are described. The evaluation of this approach was made following QA@CLEF03 criteria on a corpus of over 200,000 news in Spanish. The paper shows and discusses the results achieved by the system.

Current efforts on the semantic web are mainly focused on the creation of recommendations and standards for adding semantic descriptions to web resources. This situation represents a huge challenge to content creators that have to construct manually such descriptions, implying high costs in material and human resources. This paper presents a multi-agent system that automates partially this task, i.e. the authoring of web documents, reducing content creators labor. This system automatically extracts descriptive information from a set of documents in Spanish language, and constructs two output (web) document collections from them. The first collection is a set of meta-information descriptions based on the Dublin Core specifications. The second output is a collection of XHTML documents for human visualizing and browsing. In order to build the two output collections, the proposed multi-agent system applies several intelligent text processing approaches. This paper describes these approaches, as well as, the methodology used to encode the extracted metadata. It also reports results from processing three document collections of about 45 MB of text, including their associated resources - descriptions and hypertext - generated by the system.