Design of an extensible metadata editor based on RDF

The content of a Learning Object is frequently characterized by metadata from several standards, such as LOM, SCORM and QTI. Specialized domains require new application profiles that further complicate the task of editing the metadata of learning object since their data models are not supported by existing authoring tools. To cope with this problem we designed a metadata editor supporting multiple metadata languages, each with its own data model. It is assumed that the supported languages have an XML binding and we use RDF to create a common metadata representation, independent from the syntax of each metadata languages. The combined data model supported by the editor is defined as an ontology. Thus, the process of extending the editor to support a new metadata language is twofold: firstly, the conversion from the XML binding of the metadata language to RDF and vice-versa; secondly, the extension of the ontology to cover the new metadata model. In this paper we describe the general architecture of the editor, we explain how a typical metadata language for learning objects is represented as an ontology, and how this formalization captures all the data required to generate the graphical user interface of the editor.

Study of behavioural and neural bases of visuo-spatial working memory with an fMRI paradigm based on an n-back task

The goal of this study was to propose a new functional magnetic resonance imaging (fMRI) paradigm using a language-free adaptation of a 2-back working memory task to avoid cultural and educational bias. We additionally provide an index of the validity of the proposed paradigm and test whether the experimental task discriminates the behavioural performances of healthy participants from those of individuals with working memory deficits. Ten healthy participants and nine patients presenting working memory (WM) deficits due to acquired brain injury (ABI) performed the developed task. To inspect whether the paradigm activates brain areas typically involved in visual working memory (VWM), brain activation of the healthy participants was assessed with fMRIs. To examine the task's capacity to discriminate behavioural data, performances of the healthy participants in the task were compared with those of ABI patients. Data were analysed with GLM-based random effects procedures and t-tests. We found an increase of the BOLD signal in the specialized areas of VWM. Concerning behavioural performances, healthy participants showed the predicted pattern of more hits, less omissions and a tendency for fewer false alarms, more self-corrected responses, and faster reaction times, when compared with subjects presenting WM impairments. The results suggest that this task activates brain areas involved in VWM and discriminates behavioural performances of clinical and non-clinical groups. It can thus be used as a research methodology for behavioural and neuroimaging studies of VWM in block-design paradigms.

Hand gesture recognition system based in computer vision and machine learning

"Lecture notes in computational vision and biomechanics series, ISSN 2212-9391, vol. 19"

Vision-based portuguese sign language recognition system

Vision-based hand gesture recognition is an area of active current research in computer vision and machine learning. Being a natural way of human interaction, it is an area where many researchers are working on, with the goal of making human computer interaction (HCI) easier and natural, without the need for any extra devices. So, the primary goal of gesture recognition research is to create systems, which can identify specific human gestures and use them, for example, to convey information. For that, vision-based hand gesture interfaces require fast and extremely robust hand detection, and gesture recognition in real time. Hand gestures are a powerful human communication modality with lots of potential applications and in this context we have sign language recognition, the communication method of deaf people. Sign lan- guages are not standard and universal and the grammars differ from country to coun- try. In this paper, a real-time system able to interpret the Portuguese Sign Language is presented and described. Experiments showed that the system was able to reliably recognize the vowels in real-time, with an accuracy of 99.4% with one dataset of fea- tures and an accuracy of 99.6% with a second dataset of features. Although the im- plemented solution was only trained to recognize the vowels, it is easily extended to recognize the rest of the alphabet, being a solid foundation for the development of any vision-based sign language recognition user interface system.

Cross-language influences on new second language vocabulary acquisition: the effect of learning method

Dissertação de mestrado integrado em Psicologia

Discrimination of Brazilian propolis according to the seasoning using chemometrics and machine learning based on UV-Vis scanning data

Propolis is a chemically complex biomass produced by honeybees (Apis mellifera) from plant resins added of salivary enzymes, beeswax, and pollen. The biological activities described for propolis were also identified for donor plants resin, but a big challenge for the standardization of the chemical composition and biological effects of propolis remains on a better understanding of the influence of seasonality on the chemical constituents of that raw material. Since propolis quality depends, among other variables, on the local flora which is strongly influenced by (a)biotic factors over the seasons, to unravel the harvest season effect on the propolis chemical profile is an issue of recognized importance. For that, fast, cheap, and robust analytical techniques seem to be the best choice for large scale quality control processes in the most demanding markets, e.g., human health applications. For that, UV-Visible (UV-Vis) scanning spectrophotometry of hydroalcoholic extracts (HE) of seventy-three propolis samples, collected over the seasons in 2014 (summer, spring, autumn, and winter) and 2015 (summer and autumn) in Southern Brazil was adopted. Further machine learning and chemometrics techniques were applied to the UV-Vis dataset aiming to gain insights as to the seasonality effect on the claimed chemical heterogeneity of propolis samples determined by changes in the flora of the geographic region under study. Descriptive and classification models were built following a chemometric approach, i.e. principal component analysis (PCA) and hierarchical clustering analysis (HCA) supported by scripts written in the R language. The UV-Vis profiles associated with chemometric analysis allowed identifying a typical pattern in propolis samples collected in the summer. Importantly, the discrimination based on PCA could be improved by using the dataset of the fingerprint region of phenolic compounds ( = 280-400m), suggesting that besides the biological activities of those secondary metabolites, they also play a relevant role for the discrimination and classification of that complex matrix through bioinformatics tools. Finally, a series of machine learning approaches, e.g., partial least square-discriminant analysis (PLS-DA), k-Nearest Neighbors (kNN), and Decision Trees showed to be complementary to PCA and HCA, allowing to obtain relevant information as to the sample discrimination.

Off-line simulation inspires insight: a neurodynamics approach to efficient robot task learning

There is currently an increasing demand for robots able to acquire the sequential organization of tasks from social learning interactions with ordinary people. Interactive learning-by-demonstration and communication is a promising research topic in current robotics research. However, the efficient acquisition of generalized task representations that allow the robot to adapt to different users and contexts is a major challenge. In this paper, we present a dynamic neural field (DNF) model that is inspired by the hypothesis that the nervous system uses the off-line re-activation of initial memory traces to incrementally incorporate new information into structured knowledge. To achieve this, the model combines fast activation-based learning to robustly represent sequential information from single task demonstrations with slower, weight-based learning during internal simulations to establish longer-term associations between neural populations representing individual subtasks. The efficiency of the learning process is tested in an assembly paradigm in which the humanoid robot ARoS learns to construct a toy vehicle from its parts. User demonstrations with different serial orders together with the correction of initial prediction errors allow the robot to acquire generalized task knowledge about possible serial orders and the longer term dependencies between subgoals in very few social learning interactions. This success is shown in a joint action scenario in which ARoS uses the newly acquired assembly plan to construct the toy together with a human partner.

Hand gesture recognition system based in computer vision and machine learning: Applications on human-machine interaction

Tese de Doutoramento em Engenharia de Eletrónica e de Computadores

Collage, a Collaborative Learning Design Editor Based on Patterns

This paper introduces Collage, a high-level IMS-LD compliant authoring tool that is specialized for CSCL (Computer-Supported Collaborative Learning). Nowadays CSCL is a key trend in elearning since it highlights the importance of social interactions as an essential element of learning. CSCL is an interdisciplinary domain, which demands participatory design techniques that allow teachers to get directly involved in design activities. Developing CSCL designs using LD is a difficult task for teachers since LD is a complex technical specification and modelling collaborative characteristics can be tricky. Collage helps teachers in the process of creating their own potentially effective collaborative Learning Designs by reusing and customizing patterns, according to the requirements of a particular learning situation. These patterns, called Collaborative Learning Flow Patterns (CLFPs), represent best practices that are repetitively used by practitioners when structuring the flow of (collaborative) learning activities. An example of an LD that can be created using Collage is illustrated in the paper. Preliminary evaluation results show that teachers, with experience in CL but without LD knowledge, can successfully design real collaborative learning experiences using Collage.

Electrical neuroimaging of memory discrimination based on single-trial multisensory learning.

Multisensory experiences influence subsequent memory performance and brain responses. Studies have thus far concentrated on semantically congruent pairings, leaving unresolved the influence of stimulus pairing and memory sub-types. Here, we paired images with unique, meaningless sounds during a continuous recognition task to determine if purely episodic, single-trial multisensory experiences can incidentally impact subsequent visual object discrimination. Psychophysics and electrical neuroimaging analyses of visual evoked potentials (VEPs) compared responses to repeated images either paired or not with a meaningless sound during initial encounters. Recognition accuracy was significantly impaired for images initially presented as multisensory pairs and could not be explained in terms of differential attention or transfer of effects from encoding to retrieval. VEP modulations occurred at 100-130ms and 270-310ms and stemmed from topographic differences indicative of network configuration changes within the brain. Distributed source estimations localized the earlier effect to regions of the right posterior temporal gyrus (STG) and the later effect to regions of the middle temporal gyrus (MTG). Responses in these regions were stronger for images previously encountered as multisensory pairs. Only the later effect correlated with performance such that greater MTG activity in response to repeated visual stimuli was linked with greater performance decrements. The present findings suggest that brain networks involved in this discrimination may critically depend on whether multisensory events facilitate or impair later visual memory performance. More generally, the data support models whereby effects of multisensory interactions persist to incidentally affect subsequent behavior as well as visual processing during its initial stages.

Nondeclarative learning in children with Specific Language Impairment: Predicting regularities in the visuomotor, phonological, and cognitive domains.

Ullman (2004) suggested that Specific Language Impairment (SLI) results from a general procedural learning deficit. In order to test this hypothesis, we investigated children with SLI via procedural learning tasks exploring the verbal, motor, and cognitive domains. Results showed that compared with a Control Group, the children with SLI (a) were unable to learn a phonotactic learning task, (b) were able but less efficiently to learn a motor learning task and (c) succeeded in a cognitive learning task. Regarding the motor learning task (Serial Reaction Time Task), reaction times were longer and learning slower than in controls. The learning effect was not significant in children with an associated Developmental Coordination Disorder (DCD), and future studies should consider comorbid motor impairment in order to clarify whether impairments are related to the motor rather than the language disorder. Our results indicate that a phonotactic learning but not a cognitive procedural deficit underlies SLI, thus challenging Ullmans' general procedural deficit hypothesis, like a few other recent studies.

Forecasting a language shift based on cellular automata

Language extinction as a consequence of language shifts is a widespread social phenomenon that affects several million people all over the world today. An important task for social sciences research should therefore be to gain an understanding of language shifts, especially as a way of forecasting the extinction or survival of threatened languages, i.e., determining whether or not the subordinate language will survive in communities with a dominant and a subordinate language. In general, modeling is usually a very difficult task in the social sciences, particularly when it comes to forecasting the values of variables. However, the cellular automata theory can help us overcome this traditional difficulty. The purpose of this article is to investigate language shifts in the speech behavior of individuals using the methodology of the cellular automata theory. The findings on the dynamics of social impacts in the field of social psychology and the empirical data from language surveys on the use of Catalan in Valencia allowed us to define a cellular automaton and carry out a set of simulations using that automaton. The simulation results highlighted the key factors in the progression or reversal of a language shift and the use of these factors allowed us to forecast the future of a threatened language in a bilingual community.

New kernel functions and learning methods for text and data mining

Recent advances in machine learning methods enable increasingly the automatic construction of various types of computer assisted methods that have been difficult or laborious to program by human experts. The tasks for which this kind of tools are needed arise in many areas, here especially in the fields of bioinformatics and natural language processing. The machine learning methods may not work satisfactorily if they are not appropriately tailored to the task in question. However, their learning performance can often be improved by taking advantage of deeper insight of the application domain or the learning problem at hand. This thesis considers developing kernel-based learning algorithms incorporating this kind of prior knowledge of the task in question in an advantageous way. Moreover, computationally efficient algorithms for training the learning machines for specific tasks are presented. In the context of kernel-based learning methods, the incorporation of prior knowledge is often done by designing appropriate kernel functions. Another well-known way is to develop cost functions that fit to the task under consideration. For disambiguation tasks in natural language, we develop kernel functions that take account of the positional information and the mutual similarities of words. It is shown that the use of this information significantly improves the disambiguation performance of the learning machine. Further, we design a new cost function that is better suitable for the task of information retrieval and for more general ranking problems than the cost functions designed for regression and classification. We also consider other applications of the kernel-based learning algorithms such as text categorization, and pattern recognition in differential display. We develop computationally efficient algorithms for training the considered learning machines with the proposed kernel functions. We also design a fast cross-validation algorithm for regularized least-squares type of learning algorithm. Further, an efficient version of the regularized least-squares algorithm that can be used together with the new cost function for preference learning and ranking tasks is proposed. In summary, we demonstrate that the incorporation of prior knowledge is possible and beneficial, and novel advanced kernels and cost functions can be used in algorithms efficiently.

A Case-based reasoning approach to validate grammatical gender and number agreement in Spanish language

Across Latin America 420 indigenous languages are spoken. Spanish is considered a second language in indigenous communities and is progressively introduced in education. However, most of the tools to support teaching processes of a second language have been developed for the most common languages such as English, French, German, Italian, etc. As a result, only a small amount of learning objects and authoring tools have been developed for indigenous people considering the specific needs of their population. This paper introduces Multilingual–Tiny as a web authoring tool to support the virtual experience of indigenous students and teachers when they are creating learning objects in indigenous languages or in Spanish language, in particular, when they have to deal with the grammatical structures of Spanish. Multilingual–Tiny has a module based on the Case-based Reasoning technique to provide recommendations in real time when teachers and students write texts in Spanish. An experiment was performed in order to compare some local similarity functions to retrieve cases from the case library taking into account the grammatical structures. As a result we found the similarity function with the best performance

Kernel-Based Ranking. Methods for Learning and Performance Estimation

Machine learning provides tools for automated construction of predictive models in data intensive areas of engineering and science. The family of regularized kernel methods have in the recent years become one of the mainstream approaches to machine learning, due to a number of advantages the methods share. The approach provides theoretically well-founded solutions to the problems of under- and overfitting, allows learning from structured data, and has been empirically demonstrated to yield high predictive performance on a wide range of application domains. Historically, the problems of classification and regression have gained the majority of attention in the field. In this thesis we focus on another type of learning problem, that of learning to rank. In learning to rank, the aim is from a set of past observations to learn a ranking function that can order new objects according to how well they match some underlying criterion of goodness. As an important special case of the setting, we can recover the bipartite ranking problem, corresponding to maximizing the area under the ROC curve (AUC) in binary classification. Ranking applications appear in a large variety of settings, examples encountered in this thesis include document retrieval in web search, recommender systems, information extraction and automated parsing of natural language. We consider the pairwise approach to learning to rank, where ranking models are learned by minimizing the expected probability of ranking any two randomly drawn test examples incorrectly. The development of computationally efficient kernel methods, based on this approach, has in the past proven to be challenging. Moreover, it is not clear what techniques for estimating the predictive performance of learned models are the most reliable in the ranking setting, and how the techniques can be implemented efficiently. The contributions of this thesis are as follows. First, we develop RankRLS, a computationally efficient kernel method for learning to rank, that is based on minimizing a regularized pairwise least-squares loss. In addition to training methods, we introduce a variety of algorithms for tasks such as model selection, multi-output learning, and cross-validation, based on computational shortcuts from matrix algebra. Second, we improve the fastest known training method for the linear version of the RankSVM algorithm, which is one of the most well established methods for learning to rank. Third, we study the combination of the empirical kernel map and reduced set approximation, which allows the large-scale training of kernel machines using linear solvers, and propose computationally efficient solutions to cross-validation when using the approach. Next, we explore the problem of reliable cross-validation when using AUC as a performance criterion, through an extensive simulation study. We demonstrate that the proposed leave-pair-out cross-validation approach leads to more reliable performance estimation than commonly used alternative approaches. Finally, we present a case study on applying machine learning to information extraction from biomedical literature, which combines several of the approaches considered in the thesis. The thesis is divided into two parts. Part I provides the background for the research work and summarizes the most central results, Part II consists of the five original research articles that are the main contribution of this thesis.