Automatic learning for the classification of chemical reactions and in statistical thermodynamics

This Thesis describes the application of automatic learning methods for a) the classification of organic and metabolic reactions, and b) the mapping of Potential Energy Surfaces(PES). The classification of reactions was approached with two distinct methodologies: a representation of chemical reactions based on NMR data, and a representation of chemical reactions from the reaction equation based on the physico-chemical and topological features of chemical bonds. NMR-based classification of photochemical and enzymatic reactions. Photochemical and metabolic reactions were classified by Kohonen Self-Organizing Maps (Kohonen SOMs) and Random Forests (RFs) taking as input the difference between the 1H NMR spectra of the products and the reactants. The development of such a representation can be applied in automatic analysis of changes in the 1H NMR spectrum of a mixture and their interpretation in terms of the chemical reactions taking place. Examples of possible applications are the monitoring of reaction processes, evaluation of the stability of chemicals, or even the interpretation of metabonomic data. A Kohonen SOM trained with a data set of metabolic reactions catalysed by transferases was able to correctly classify 75% of an independent test set in terms of the EC number subclass. Random Forests improved the correct predictions to 79%. With photochemical reactions classified into 7 groups, an independent test set was classified with 86-93% accuracy. The data set of photochemical reactions was also used to simulate mixtures with two reactions occurring simultaneously. Kohonen SOMs and Feed-Forward Neural Networks (FFNNs) were trained to classify the reactions occurring in a mixture based on the 1H NMR spectra of the products and reactants. Kohonen SOMs allowed the correct assignment of 53-63% of the mixtures (in a test set). Counter-Propagation Neural Networks (CPNNs) gave origin to similar results. The use of supervised learning techniques allowed an improvement in the results. They were improved to 77% of correct assignments when an ensemble of ten FFNNs were used and to 80% when Random Forests were used. This study was performed with NMR data simulated from the molecular structure by the SPINUS program. In the design of one test set, simulated data was combined with experimental data. The results support the proposal of linking databases of chemical reactions to experimental or simulated NMR data for automatic classification of reactions and mixtures of reactions. Genome-scale classification of enzymatic reactions from their reaction equation. The MOLMAP descriptor relies on a Kohonen SOM that defines types of bonds on the basis of their physico-chemical and topological properties. The MOLMAP descriptor of a molecule represents the types of bonds available in that molecule. The MOLMAP descriptor of a reaction is defined as the difference between the MOLMAPs of the products and the reactants, and numerically encodes the pattern of bonds that are broken, changed, and made during a chemical reaction. The automatic perception of chemical similarities between metabolic reactions is required for a variety of applications ranging from the computer validation of classification systems, genome-scale reconstruction (or comparison) of metabolic pathways, to the classification of enzymatic mechanisms. Catalytic functions of proteins are generally described by the EC numbers that are simultaneously employed as identifiers of reactions, enzymes, and enzyme genes, thus linking metabolic and genomic information. Different methods should be available to automatically compare metabolic reactions and for the automatic assignment of EC numbers to reactions still not officially classified. In this study, the genome-scale data set of enzymatic reactions available in the KEGG database was encoded by the MOLMAP descriptors, and was submitted to Kohonen SOMs to compare the resulting map with the official EC number classification, to explore the possibility of predicting EC numbers from the reaction equation, and to assess the internal consistency of the EC classification at the class level. A general agreement with the EC classification was observed, i.e. a relationship between the similarity of MOLMAPs and the similarity of EC numbers. At the same time, MOLMAPs were able to discriminate between EC sub-subclasses. EC numbers could be assigned at the class, subclass, and sub-subclass levels with accuracies up to 92%, 80%, and 70% for independent test sets. The correspondence between chemical similarity of metabolic reactions and their MOLMAP descriptors was applied to the identification of a number of reactions mapped into the same neuron but belonging to different EC classes, which demonstrated the ability of the MOLMAP/SOM approach to verify the internal consistency of classifications in databases of metabolic reactions. RFs were also used to assign the four levels of the EC hierarchy from the reaction equation. EC numbers were correctly assigned in 95%, 90%, 85% and 86% of the cases (for independent test sets) at the class, subclass, sub-subclass and full EC number level,respectively. Experiments for the classification of reactions from the main reactants and products were performed with RFs - EC numbers were assigned at the class, subclass and sub-subclass level with accuracies of 78%, 74% and 63%, respectively. In the course of the experiments with metabolic reactions we suggested that the MOLMAP / SOM concept could be extended to the representation of other levels of metabolic information such as metabolic pathways. Following the MOLMAP idea, the pattern of neurons activated by the reactions of a metabolic pathway is a representation of the reactions involved in that pathway - a descriptor of the metabolic pathway. This reasoning enabled the comparison of different pathways, the automatic classification of pathways, and a classification of organisms based on their biochemical machinery. The three levels of classification (from bonds to metabolic pathways) allowed to map and perceive chemical similarities between metabolic pathways even for pathways of different types of metabolism and pathways that do not share similarities in terms of EC numbers. Mapping of PES by neural networks (NNs). In a first series of experiments, ensembles of Feed-Forward NNs (EnsFFNNs) and Associative Neural Networks (ASNNs) were trained to reproduce PES represented by the Lennard-Jones (LJ) analytical potential function. The accuracy of the method was assessed by comparing the results of molecular dynamics simulations (thermal, structural, and dynamic properties) obtained from the NNs-PES and from the LJ function. The results indicated that for LJ-type potentials, NNs can be trained to generate accurate PES to be used in molecular simulations. EnsFFNNs and ASNNs gave better results than single FFNNs. A remarkable ability of the NNs models to interpolate between distant curves and accurately reproduce potentials to be used in molecular simulations is shown. The purpose of the first study was to systematically analyse the accuracy of different NNs. Our main motivation, however, is reflected in the next study: the mapping of multidimensional PES by NNs to simulate, by Molecular Dynamics or Monte Carlo, the adsorption and self-assembly of solvated organic molecules on noble-metal electrodes. Indeed, for such complex and heterogeneous systems the development of suitable analytical functions that fit quantum mechanical interaction energies is a non-trivial or even impossible task. The data consisted of energy values, from Density Functional Theory (DFT) calculations, at different distances, for several molecular orientations and three electrode adsorption sites. The results indicate that NNs require a data set large enough to cover well the diversity of possible interaction sites, distances, and orientations. NNs trained with such data sets can perform equally well or even better than analytical functions. Therefore, they can be used in molecular simulations, particularly for the ethanol/Au (111) interface which is the case studied in the present Thesis. Once properly trained, the networks are able to produce, as output, any required number of energy points for accurate interpolations.

Learning with tangible interfaces

Trabalho apresentado no âmbito do Mestrado em Engenharia Informática, como requisito parcial para obtenção do grau de Mestre em Engenharia Informática

E-learning frameworks: a survey

In recent years the concept of eLearning Framework emerged associated with several initiatives promoted by educational organizations. These initiatives share a common goal: to create flexible learning environments by integrating heterogeneous systems already available in many educational institutions. The paper provides an introductory survey on eLearning Frameworks. It gathers information on these initiatives categorizes them and compares their features regarding a set of predefined criteria such as: architecture, business model, primary user groups, technical implementations, adopted standards, maturity and future development.

Using the Common Cartridge profile to enhance learning content interoperability

The concept of Learning Object (LO) is crucial for the standardization on eLearning. The latest LO standard from IMS Global Learning Consortium is the IMS Common Cartridge (IMS CC) that organizes and distributes digital learning content. By analyzing this new specification we considered two interoperability levels: content and communication. A common content format is the backbone of interoperability and is the basis for content exchange among eLearning systems. Communication is more than just exchanging content; it includes also accessing to specialized systems and services and reporting on content usage. This is particularly important when LOs are used for evaluation. In this paper we analyze the Common Cartridge profile based on the two interoperability levels we proposed. We detail its data model that comprises a set of derived schemata referenced on the CC schema and we explore the use of the IMS Learning Tools Interoperability (LTI) to allow remote tools and content to be integrated into a Learning Management System (LMS). In order to test the applicability of IMS CC for automatic evaluation we define a representation of programming exercises using this standard. This representation is intended to be the cornerstone of a network of eLearning systems where students can solve computer programming exercises and obtain feedback automatically. The CC learning object is automatically generated based on a XML dialect called PExIL that aims to consolidate all the data need to describe resources within the programming exercise life-cycle. Finally, we test the generated cartridge on the IMS CC online validator to verify its conformance with the IMS CC specification.

Mobile learning: recetividade e predisposição do ensino superior para a sua utilização

A evolução dos dispositivos móveis e a mudança de paradigma educacional, permitiu o surgimento de um novo conceito no processo de ensino e aprendizagem, o mobile learning. O mobile learning pode ser visto como um conceito multidisciplinar, dependendo da perspetiva de cada autor, pois ainda não existe um consenso em relação à definição do conceito. No entanto, todos os autores concordam que o mobile learning consiste na aquisição de conhecimento ou competência através do uso de tecnologias móveis, em qualquer lugar e momento. A presente investigação, de natureza exploratória, pretendeu estudar a receptividade e predisposição dos estudantes e docentes do ensino superior para com a utilização do mobile learning, uma vez que o ensino superior parece ser o ambiente ideal para a realização deste estudo. Por um lado, devido à democratização dos dispositivos móveis, por outro, porque o Instituto Politécnico do Porto pretende vir a implementar um projeto de mobile learning, enquadrado no e- IPP. Deste modo, para a concretização desta investigação, foi realizada uma revisão bibliográfica exaustiva que serviu de base de sustentação para todo o trabalho, complementada com um questionário, de forma a dar resposta às questões de investigação. Depois de recolhidos todos os resultados obtidos através do questionário, procedeu-se à análise e discussão mesmos, bem como às respectivas conclusões.

A Learning Algorithm and System Approach to Address Exceptional Events in Domestic Consumption Management

The integration of the Smart Grid concept into the electric grid brings to the need for an active participation of small and medium players. This active participation can be achieved using decentralized decisions, in which the end consumer can manage loads regarding the Smart Grid needs. The management of loads must handle the users’ preferences, wills and needs. However, the users’ preferences, wills and needs can suffer changes when faced with exceptional events. This paper proposes the integration of exceptional events into the SCADA House Intelligent Management (SHIM) system developed by the authors, to handle machine learning issues in the domestic consumption context. An illustrative application and learning case study is provided in this paper.

Definição do conceito de blended learning. Proposta metodológica no quadro da terminologia de base conceptual

Sustentado numa abordagem conceptual da Terminologia, o trabalho de investigação que a seguir se desenvolve visa propor um cenário de resposta à questão: como se define o blended learning no contexto educativo de Ensino Superior Pós-Bolonha? A necessidade de compreender, delimitar e harmonizar o conceito de blended learning no panorama actual do Ensino Superior, resulta do pressuposto de que muito embora proliferem descrições de práticas e modelos para a sua operacionalização de reconhecida qualidade - tal como sucede com outros conceitos sob alçada da educação a distância - a reflexão teórica é ainda insuficiente. Com efeito, para responder à questão supra-colocada, propõe-se o desenho de uma metodologia para construção de uma definição intensional do conceito de blended learning que herde, subsuma e melhore o conhecimento existente, identificado através da análise de texto para fins onomasiológicos e de um processo de elicitação de conhecimento tácito e de negociação discursiva junto de sujeitos especializados. A proposta de desenho metodológico que neste trabalho de esboça escora-se globalmente em três etapas: (1) etapa exploratória do domínio-objecto de estudo; (2) etapa de análise onamasiológica de evidência textual e discursiva; (3) etapa de modelização e de validação de resultados. Pretende-se, em primeiro lugar, através do estudo do espaço conceptual das modalidades de educação que se situam no continuum presença-distância sistematizar e ordenar as visões analisadas, propondo representações de educação presencial, educação a distância, e-learning, educação online, aprendizagem enriquecida por tecnologias e ainda de outras modalidades emergentes. As representações assumem o carácter de proposta aberta e decorrem da necessidade de uma primeira ordenação no sentido de topografrar, delineando - a um nível macro - o possível lugar do conceito de blended learning naquele cenário. Num segundo momento, aprofunda-se e circunscreve-se a análise, depurando a evidência observada, agora reduzida a um conjunto de contextos ricos em informação conceptual – um corpus escrito e oral de definições e descrições de blended learning – identificando candidatas a características essenciais e candidatas a características acidentais. Num terceiro momento, procede-se à modelização do conhecimento, encapsulando-a numa proposta de definição sujeita a um processo iterativo e reflexivo, constituído por um conjunto de ciclos de investigação-acção, os quais reflectem a sequência de interacções entre o terminólogo e sujeitos especializados. Defender-se-á que a experimentação deste desenho revela a produtividade de uma sequência cíclica entre a análise textual e discursiva para fins onomasiológicos, a interacção colaborativa e a introspecção. Por outras palavras, embora a natureza do estudo realizado não permita a generalização, para além da relação diádica de mediação que o terminológo estabelece com o especialista, defende-se a produtividade de um procedimento de acção-reflexão autónomo, solitário e introspectivo, no âmbito da qual o terminólogo se afirma como sujeito conceptualizador, decisor e interventor. Resultam deste percurso uma proposta de definição e de descrição de blended learning em língua portuguesa que acreditamos poder servir diferentes actores da comunidade académica envolvida neste domínio de especialidade.

“Learning-by-producing” in the fashion value chain

Proceedings da AUTEX 2015, Bucareste, Roménia.

Open Learning Network : the evidence of OER impact

Much of the initial work on Open Educational Resources (OER) has inevitably concentrated on how to produce the resources themselves and to establish the idea in the community. It is now eight years since the term OER was first used and more than ten years since the concept of open content was described and a greater focus is now emerging on the way in which OER can influence policy and change the way in which educational systems help people learn. The Open University UK and Carnegie Mellon University are working in partnership on the OLnet (Open Learning Network), funded by The William and Flora Hewlett Foundation with the aims to search out the evidence for use and reuse of OER and to establish a network for information sharing about research in the field. This means both gathering evidence and developing approaches for how to research and understand ways to learn in a more open world, particularly linked to OER, but also looking at other influences.

Towards a Social Learning Space for Open Educational Resources

We identify a number of meanings of "Open", as part of the motivating rationale for a social media space tuned for learning, called SocialLearn. We discuss why online social learning seems to be emerging so strongly at this point, explore features of social learning, and identify some of the dimensions that we believe characterize the social learning design space, before describing the emerging design concept and implementation.

Hacia una definición inclusiva del e-learning

El estudio tiene como objetivo la construcción de una definición actual e integradora del concepto de e-learning, que sea aceptada por la mayor parte de la comunidad científica y que sirva como referente por los estudiosos y profesionales de esta temática.

Towards and inclusive definition of e-learning

L'estudi té com a objectiu la construcció d'una definició actual i integradora del concepte d'e-learning, que sigui acceptada per la major part de la comunitat científica i que serveixi com a referent pels estudiosos i professionals d'aquesta temàtica.

Cap a una definició inclusiva de l'e-learning

Peer-reviewed

Transapical aortic valve implantation without angiography: proof of concept.

Background: Cardiac computed tomographic scans, coronary angiograms, and aortographies are routinely performed in transcatheter heart valve therapies. Consequently, all patients are exposed to multiple contrast injections with a following risk of nephrotoxicity and postoperative renal failure. The transapical aortic valve implantation without angiography can prevent contrast-related complications. Methods: Between November 2008 and November 2009, 30 consecutive high-risk patients (16 female, 53.3%) underwent transapical aortic valve implantation without angiography. The landmarks identification, the stent-valve positioning, and the postoperative control were routinely performed under transesophageal echocardiogram and fluoroscopic visualization without contrast injections. Results: Mean age was 80.1 +/- 8.7 years. Mean valve gradient, aortic orifice area, and ejection fraction were 60.3 +/- 20.9 mm Hg, 0.7 +/- 0.16 cm(2), and 0.526 +/- 0.128, respectively. Risk factors were pulmonary hypertension (60%), peripheral vascular disease (70%), chronic pulmonary disease (50%), previous cardiac surgery (13.3%), and chronic renal insufficiency (40%) (mean blood creatinine and urea levels: 96.8 +/- 54 mu g/dL and 8.45 +/- 5.15 mmol/L). Average European System for Cardiac Operative Risk Evaluation was 32.2 +/- 13.3%. Valve deployment in the ideal landing zone was 96.7% successful and valve embolization occurred once. Thirty-day mortality was 10% (3 patients). Causes of death were the following: intraoperative ventricular rupture (conversion to sternotomy), right ventricular failure, and bilateral pneumonia. Stroke occurred in one patient at postoperative day 9. Renal failure (postoperative mean blood creatinine and urea levels: 91.1 +/- 66.8 mu g/dL and 7.27 +/- 3.45 mmol/L), myocardial infarction, and atrioventricular block were not detected. Conclusions: Transapical aortic valve implantation without angiography requires a short learning curve and can be performed routinely by experienced teams. Our report confirms that this procedure is feasible and safe, and provides good results with low incidence of postoperative renal disorders. (Ann Thorac Surg 2010; 89: 1925-33) (C) 2010 by The Society of Thoracic Surgeons

Learning and adaptation in physical heterogeneous teams of robots

In this paper we present a novel approach to assigning roles to robots in a team of physical heterogeneous robots. Its members compete for these roles and get rewards for them. The rewards are used to determine each agent’s preferences and which agents are better adapted to the environment. These aspects are included in the decision making process. Agent interactions are modelled using the concept of an ecosystem in which each robot is a species, resulting in emergent behaviour of the whole set of agents. One of the most important features of this approach is its high adaptability. Unlike some other learning techniques, this approach does not need to start a whole exploitation process when the environment changes. All this is exemplified by means of experiments run on a simulator. In addition, the algorithm developed was applied as applied to several teams of robots in order to analyse the impact of heterogeneity in these systems