José A. Macías

Evolution of functional link networks

This paper addresses the genetic design of functional link networks (FLN). FLN are high-order perceptrons (HOP) without hidden units. Despite their linear nature, FLN can capture nonlinear input-output relationships, provided that they are fed with an adequate set of polynomial inputs, which are constructed out of the original input attributes. Given this set, it turns out to be very simple to train the network, as compared with a multilayer perceptron (MLP). However finding the optimal subset of units is a difficult problem because of its nongradient nature and the large number of available units, especially for high degrees. Some constructive growing methods have been proposed to address this issue, Here, we rely on the global search capabilities of a genetic algorithm to scan the space of subsets of polynomial units, which is plagued by a host of local minima. By contrast, the quadratic error function of each individual FLN has only one minimum, which makes fitness evaluation practically noiseless. We find that surprisingly simple FLN compare favorably with other more complex architectures derived by means of constructive and evolutionary algorithms on some UCI benchmark data sets. Moreover, our models are especially amenable to interpretation, due to an incremental approach that penalizes complex architectures and starts with a pool of single-attribute FLN.

Customization of Web applications through an intelligent environment exploiting logical interface descriptions

Customization of Web-based applications is often considered a designer skill rather than an end-user need. However, there is an ongoing shift to end-user-centred technology, and even users with poor or no skill in Web-based languages may feel the need to customize Web applications according to their preferences. Although Web authoring environments have an increasing number of features, the challenge of providing end-users with the ability to easily customize entire Web applications still remains unsolved. In this paper, we propose an intelligent approach to customizing Web-based applications. Customizations rules are automatically inferred by the system from changes that users supply as examples. They remain as long-term knowledge that can be applied to support future interactions, thus minimizing the amount of authoring that end-users need to do for this purpose. In order to better understand the implications of the users modifications, they are analysed using the logical descriptions of the corresponding Web pages.

Enhancing Project-Based Learning in Software Engineering Lab Teaching Through an E-Portfolio Approach

Project-based learning is one of the main successful student-centered pedagogies broadly used in computing science courses. However, this approach can be insufficient when dealing with practical subjects that implicitly require many deliverables and a great deal of feedback and organizational resources. In this paper, a worked e-portfolio is presented as an approach to improve the teaching/learning and evaluation processes in project-based learning environments needing considerable resources. To validate this approach, a practical project-based software engineering course supported by a Moodle-based e-portfolio was designed and taught. The results obtained corroborated the effectiveness of the e-portfolio in practical software engineering teaching; this approach can be extended to similar subjects in other studies and/or curricula.

Ontology-Based Retrieval of Human Speech

As part of the general growth and diversification of media in different modalities, the presence of information in the form of human speech in the world-wide body of digital content is becoming increasingly significant, in terms of both volume and value. We present a semantic- based search model for human speech corpora, stressing the search for meanings rather than words. Our framework embraces the complete recognition/retrieval cycle, from word spotting to semantic annotation, query processing, and search result presentation.

A model-driven approach to building modern Semantic Web-Based User Interfaces

The Semantic Web has widely spread in the last 10 years as a suitable web platform to support semantics and expressive information seeking. However, one of the main problems with this paradigm is still the representation and manipulation of ontologies as well as the complex relationships that they implicitly represent. Actually, this remains a challenge when unskilled users have to deal with this abstract representation in order to carry out daily solving-problem activities (e.g., designing web applications based on ontologies). This probably made the Semantic Web to decrease in popularity, also being commercially unsupported and overcame by recent technologies and services based on the Web 2.0, the emerging end-user-focused web concept. All in all, the specification of Model-Based User Interfaces fits very well to both paradigms. Accordingly, the aim of this work is to provide new ways of modeling user interfaces based on semantic models that better fit the domain problem. At the same time, we think of exploiting interactive features through current and modern end-user programming elements based on the Web 2.0, finally contributing to an architecture that supports higher interactive end-user interfaces on the web.

Providing end-user facilities to simplify ontology-driven web application authoring

Generally speaking, emerging web-based technologies are mostly intended for professional developers. They pay poor attention to users who have no programming abilities but need to customize software applications. At some point, such needs force end-users to act as designers in various aspects of software authoring and development. Every day, more new computing-related professionals attempt to create and modify existing applications in order to customize web-based artifacts that will help them carry out their daily tasks. In general they are domain experts rather than skilled software designers, and new authoring mechanisms are needed in order that they can accomplish their tasks properly. The work we present is an effort to supply end-users with easy mechanisms for authoring web-based applications. To complement this effort, we present a user study showing that it is possible to carry out a trade-off between expressiveness and ease of use in order to provide end-users with authoring facilities.

Dynamic web page authoring by example using ontology-based domain knowledge

Authoring dynamic web pages is an inherently difficult task. We present DESK, an interactive authoring tool that allows the customization of dynamic page generation procedures with no a-priori tool-specific skill requirements from authors. Our approach consists of combining Programming By Example (PBE) techniques with an ontology-based representation of knowledge displayed in web pages. DESK acts as a client-side complement of a dynamic web page generation system, PEGASUS, which generates HTML pages from a formally structured domain model and an abstract presentation model. Authorized users can modify the internal presentation model by editing the generated HTML pages with DESK in a WYSIWYG environment. DESK keeps track of all users actions and exploits the explicitly represented domain semantics to enhance the power of PBE techniques

Feasible database querying using a visual end-user approach

Querying databases is a common daily task carried out by a great deal of end-users who do not have specific skills in SQL language. Today, most of the database interaction is achieved by means of query interfaces provided by the database environment. However, most of these interfaces suffer from expressive limitations, since they are mostly based on metaphors that drastically restrict the expressiveness of the SQL language that is generated and executed in the background. In this paper, we present a visual interaction language and tool focused on easily querying databases by end-users. We make no assumption on the level of the users experience with query languages, as our visual metaphor is intended for querying databases by unskilled end-users and also leveraging the restriction on the expressiveness of the queries created by them. We also report on some late braking results obtained by an experiment carried out with real users.

A generic presentation modeling system for adaptive web-based instructional applications

We propose a generic presentation system for adaptive educational hypermedia that is highly independent from domain knowledge representation and application state management. Generality is achieved by providing a framework for the definition of ontologies that best fit specific domains and/or authors. Presentations are described in terms of ontology object classes and relations. An explicit presentation model, separate from course contents, is used to provide course designers with extensive control over the generation of all aspects of presentation, at a moderate development cost.

Interactive Design of Adaptive Courses

The ability of educational applications to adapt to individual students has been a sought-after feature for more than a decade. The difficulty of developing this kind of software and the lack of adequate tools that facilitate this endeavour have hindered the participation of instructors and teachers in the elaboration of adaptive applications. In this paper we describe an authoring tool, ATLAS, that aims at combining expressive power and ease of use in the design of adaptive web courses. ATLAS allows the fully interactive construction of courses that adapt automatically to the student’s characteristics and her/his behaviour while taking the course. The designer interacts with the tool by using an intuitive visual language based on the direct manipulation of elements involved in the course. The tool takes care of the transition between a teacher’s understanding of the course and the representation model of the underlying system.