Azucena Montes Rendón

Compilation of symbolic knowledge and integration with numeric knowledge using hybrid systems

The development of Artificial Intelligence (AI) research has followed mainly two directions: the use of symbolic and connectionist (artificial neural networks) methods. These two approaches have been applied separately in the solution of problems that require tasks of knowledge acquisition and learning. We present the results of implementing a Neuro-Symbolic Hybrid System (NSHS) that allows unifying these two types of knowledge representation. For this, we have developed a compiler or translator of symbolic rules which takes as an input a group of rules of the type IF ... THEN..., converting them into a connectionist representation. Obtained the compiled artificial neural network this is used as an initial neural network in a learning process that will allow the “refinement” of the knowledge. To prove the refinement of the hybrid approach, we carried out a group of tests that show that it is possible to improve in a connectionist way the symbolic knowledge.

NLP Methodology as Guidance and Verification of the Data Mining of Survey ENSANUT 2012

Data Mining represents the cutting edge when we think about extracting information; however it always implicates a considerable spent provided that it needs “structured data”. Following this idea, text mining appears in the horizon, as a little spent, reliable alternative. It is able to provide meaningful expert information without the availability of plenty of resources, all we need is a fair big (real big) corpus of text in order to conduct a research on almost every topic. By themselves, both approaches provide valuable information at the end, nevertheless what would happen if both processes were linked in a way that one approach’s results could be verify by the result of a second process? With this idea on mind we are relaying on one hypothesis this is possible to generate a bound between both mining process and using them back and forth to verify one another. Hence, we describe thoroughly both methodologies making a special emphasis on mentioning those phases which have a propensity to establish a strong bound between them. We found that bound in the fact that once a Natural Language Processing has been performed on the chosen corpora what we got as an output is a list of meaningful nouns which can be used as features that will guide in a reliable way a data mining process.

An On-the---Flight Proxy System for Multi-format Web Content Transcoding for Mobile Devices

One of the trends in current information society is the visualization of Web sites using mobile devices. At present time, there are millions of mobile devices that are used by the owner to navigate in the web. However, not all the information placed in the web is appropriately adapted to be visualized in mobile devices. As a result, not all functionalities can be exploded by the users. To solve this situation, we need mechanisms to allow users of mobile devices to visualize Web sites independently if the Web sites were specifically designed or not for these types of devices. The problem of this scenario is that the contents of the Web site need to be adapted to the limitations of each mobile device. In this paper, an on-the-flight Web content transcoding mechanism is presented as solution to this problem. The mechanism proposes the inclusion of a new HTTP header in the HTTP package to describe a transcoding format command. This header is generated by a client-side Proxy named GAP, which runs on mobile devices with Windows CE operating system (Pocket PC, Smartphone, among others). The transcoding header is interpreted by a server-side Proxy, named GAT MT, that converts Web documents to WML, XHTML-MP, PDF, plain text, PostScript, XML and HTML adapted for mobile devices. To do this, the mechanism uses XSL technologies (XSLT and XSL-FO) to define Stylesheets according to a specific device.

A Method for Automatic Text Categorization Using Word Sense Disambiguation

At present time, Information plays a relevant role in current societies. In this context, Internet is one of the most extended mechanisms to communicate and distribute information around the word. Today, due to the extremely large number of information sources, automatic mechanisms are needed to filter the information that could be useful for each user. However, one of the problems that the usual techniques of automatic text categorization have not been able to handle is polysemy (words with two o more senses). In this paper, we have faced this problem by proposing a semantic analyzer for the automatic categorization of texts in Spanish. Context exploration techniques were used as a key mechanism for guiding the disambiguation process. A specific lexical database and its existing semantic relations fulfilled the objective of appropriately categorizing the analyzed text. To validate this analyzer, a tool was developed that classifies web pages by semantic sense. We present performance results for this classifier. Finally, a comparison with four other classification tools is reported.

Using Web Pages Accessible Design for the Correct Web Visualization

The objective of this work is to use the recommendations for the accessible design of Web pages and applying them for the correct visualization of Web resources on mobile devices. A small study was made taking in account these considerations, obtaining that are feasible to apply these recommendations or practices of good design for their use on mobile devices. Also a proposal appears to solve the problem of correct Web pages visualization on mobile devices.

Automatic description of static images in natural language

In this paper, the description of an image in natural language is carried out. The main idea is that from an image, with objects without movement, it is possible to obtain phrases in Spanish describing the position among the objects. In order to put this description into effect, we place ourselves in a theoretical model in which a cognitive-semantic analysis of linguistic units such as the prepositions sobre (on), en (in), entre (between) and the verb tocar (to touch) is realized. This analysis will allow to establish rules which will determine the relationship or position among the objects.