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.

Fast Learning by Bounding Likelihoods in Sigmoid Type Belief Networks

Sigmoid type belief networks, a class of probabilistic neural networks, provide a natural framework for compactly representing probabilistic information in a variety of unsupervised and supervised learning problems. Often the parameters used in these networks need to be learned from examples. Unfortunately, estimating the parameters via exact probabilistic calculations (i.e, the EM-algorithm) is intractable even for networks with fairly small numbers of hidden units. We propose to avoid the infeasibility of the E step by bounding likelihoods instead of computing them exactly. We introduce extended and complementary representations for these networks and show that the estimation of the network parameters can be made fast (reduced to quadratic optimization) by performing the estimation in either of the alternative domains. The complementary networks can be used for continuous density estimation as well.

People Recognition in Image Sequences by Supervised Learning

We describe a system that learns from examples to recognize people in images taken indoors. Images of people are represented by color-based and shape-based features. Recognition is carried out through combinations of Support Vector Machine classifiers (SVMs). Different types of multiclass strategies based on SVMs are explored and compared to k-Nearest Neighbors classifiers (kNNs). The system works in real time and shows high performance rates for people recognition throughout one day.

With a little help from our friends: how system dynamics and soft OR can learn from each other

At its inception, the paradigm of system dynamics was deliberately made distinct from that of OR. Yet developments in soft OR now have much in common with current system dynamics modeling practice. This article briefly traces the parallel development of system dynamics and soft OR, and argues that a dialogue between the two would be mutually rewarding. to support this claim, examples of soft OR tools are described along with some of the field’s philosophical grounding and current issues. Potential benefits resulting from a dialogue are explored, with particular emphasis on the methodological framework of system dynamics and the need for a complementarist approach. The article closes with some suggestions on how to begin learning from.

Adaptation to climate change in megacities of Latin America: Regional Learning Network of the research project ClimateAdaptationSantiago (CAS)

The main objective of this publication is to document the current state of urban climate change adaptation practice in Latin America. It is a summary of the three workshops of the Regional Learning Network that was set up under the ClimateAdaptationSantiago project (CAS), encompassing six large Latin American cities (Bogotá, Buenos Aires, Lima, Mexico City, São Paulo and Santiago). It aims to synthesize information on the manifestations and impacts of climate change in those Latin American cities that participated in the network, and above all, governance in the form of concrete actions. The publication is based on information obtained from the participants in the three workshops, but also includes additional scientific input and reflections by the editors. All of this information makes a major contribution to highlighting the different paths these six cities are pursuing in response to climate change. To that end, the publication discusses the various courses of action on climate change adaptation, with the aim of learning from these cases and highlighting practical examples.

Contributions to Bayesian network learning with applications to neuroscience

Neuronal morphology is a key feature in the study of brain circuits, as it is highly related to information processing and functional identification. Neuronal morphology affects the process of integration of inputs from other neurons and determines the neurons which receive the output of the neurons. Different parts of the neurons can operate semi-independently according to the spatial location of the synaptic connections. As a result, there is considerable interest in the analysis of the microanatomy of nervous cells since it constitutes an excellent tool for better understanding cortical function. However, the morphologies, molecular features and electrophysiological properties of neuronal cells are extremely variable. Except for some special cases, this variability makes it hard to find a set of features that unambiguously define a neuronal type. In addition, there are distinct types of neurons in particular regions of the brain. This morphological variability makes the analysis and modeling of neuronal morphology a challenge. Uncertainty is a key feature in many complex real-world problems. Probability theory provides a framework for modeling and reasoning with uncertainty. Probabilistic graphical models combine statistical theory and graph theory to provide a tool for managing domains with uncertainty. In particular, we focus on Bayesian networks, the most commonly used probabilistic graphical model. In this dissertation, we design new methods for learning Bayesian networks and apply them to the problem of modeling and analyzing morphological data from neurons. The morphology of a neuron can be quantified using a number of measurements, e.g., the length of the dendrites and the axon, the number of bifurcations, the direction of the dendrites and the axon, etc. These measurements can be modeled as discrete or continuous data. The continuous data can be linear (e.g., the length or the width of a dendrite) or directional (e.g., the direction of the axon). These data may follow complex probability distributions and may not fit any known parametric distribution. Modeling this kind of problems using hybrid Bayesian networks with discrete, linear and directional variables poses a number of challenges regarding learning from data, inference, etc. In this dissertation, we propose a method for modeling and simulating basal dendritic trees from pyramidal neurons using Bayesian networks to capture the interactions between the variables in the problem domain. A complete set of variables is measured from the dendrites, and a learning algorithm is applied to find the structure and estimate the parameters of the probability distributions included in the Bayesian networks. Then, a simulation algorithm is used to build the virtual dendrites by sampling values from the Bayesian networks, and a thorough evaluation is performed to show the model’s ability to generate realistic dendrites. In this first approach, the variables are discretized so that discrete Bayesian networks can be learned and simulated. Then, we address the problem of learning hybrid Bayesian networks with different kinds of variables. Mixtures of polynomials have been proposed as a way of representing probability densities in hybrid Bayesian networks. We present a method for learning mixtures of polynomials approximations of one-dimensional, multidimensional and conditional probability densities from data. The method is based on basis spline interpolation, where a density is approximated as a linear combination of basis splines. The proposed algorithms are evaluated using artificial datasets. We also use the proposed methods as a non-parametric density estimation technique in Bayesian network classifiers. Next, we address the problem of including directional data in Bayesian networks. These data have some special properties that rule out the use of classical statistics. Therefore, different distributions and statistics, such as the univariate von Mises and the multivariate von Mises–Fisher distributions, should be used to deal with this kind of information. In particular, we extend the naive Bayes classifier to the case where the conditional probability distributions of the predictive variables given the class follow either of these distributions. We consider the simple scenario, where only directional predictive variables are used, and the hybrid case, where discrete, Gaussian and directional distributions are mixed. The classifier decision functions and their decision surfaces are studied at length. Artificial examples are used to illustrate the behavior of the classifiers. The proposed classifiers are empirically evaluated over real datasets. We also study the problem of interneuron classification. An extensive group of experts is asked to classify a set of neurons according to their most prominent anatomical features. A web application is developed to retrieve the experts’ classifications. We compute agreement measures to analyze the consensus between the experts when classifying the neurons. Using Bayesian networks and clustering algorithms on the resulting data, we investigate the suitability of the anatomical terms and neuron types commonly used in the literature. Additionally, we apply supervised learning approaches to automatically classify interneurons using the values of their morphological measurements. Then, a methodology for building a model which captures the opinions of all the experts is presented. First, one Bayesian network is learned for each expert, and we propose an algorithm for clustering Bayesian networks corresponding to experts with similar behaviors. Then, a Bayesian network which represents the opinions of each group of experts is induced. Finally, a consensus Bayesian multinet which models the opinions of the whole group of experts is built. A thorough analysis of the consensus model identifies different behaviors between the experts when classifying the interneurons in the experiment. A set of characterizing morphological traits for the neuronal types can be defined by performing inference in the Bayesian multinet. These findings are used to validate the model and to gain some insights into neuron morphology. Finally, we study a classification problem where the true class label of the training instances is not known. Instead, a set of class labels is available for each instance. This is inspired by the neuron classification problem, where a group of experts is asked to individually provide a class label for each instance. We propose a novel approach for learning Bayesian networks using count vectors which represent the number of experts who selected each class label for each instance. These Bayesian networks are evaluated using artificial datasets from supervised learning problems. Resumen La morfología neuronal es una característica clave en el estudio de los circuitos cerebrales, ya que está altamente relacionada con el procesado de información y con los roles funcionales. La morfología neuronal afecta al proceso de integración de las señales de entrada y determina las neuronas que reciben las salidas de otras neuronas. Las diferentes partes de la neurona pueden operar de forma semi-independiente de acuerdo a la localización espacial de las conexiones sinápticas. Por tanto, existe un interés considerable en el análisis de la microanatomía de las células nerviosas, ya que constituye una excelente herramienta para comprender mejor el funcionamiento de la corteza cerebral. Sin embargo, las propiedades morfológicas, moleculares y electrofisiológicas de las células neuronales son extremadamente variables. Excepto en algunos casos especiales, esta variabilidad morfológica dificulta la definición de un conjunto de características que distingan claramente un tipo neuronal. Además, existen diferentes tipos de neuronas en regiones particulares del cerebro. La variabilidad neuronal hace que el análisis y el modelado de la morfología neuronal sean un importante reto científico. La incertidumbre es una propiedad clave en muchos problemas reales. La teoría de la probabilidad proporciona un marco para modelar y razonar bajo incertidumbre. Los modelos gráficos probabilísticos combinan la teoría estadística y la teoría de grafos con el objetivo de proporcionar una herramienta con la que trabajar bajo incertidumbre. En particular, nos centraremos en las redes bayesianas, el modelo más utilizado dentro de los modelos gráficos probabilísticos. En esta tesis hemos diseñado nuevos métodos para aprender redes bayesianas, inspirados por y aplicados al problema del modelado y análisis de datos morfológicos de neuronas. La morfología de una neurona puede ser cuantificada usando una serie de medidas, por ejemplo, la longitud de las dendritas y el axón, el número de bifurcaciones, la dirección de las dendritas y el axón, etc. Estas medidas pueden ser modeladas como datos continuos o discretos. A su vez, los datos continuos pueden ser lineales (por ejemplo, la longitud o la anchura de una dendrita) o direccionales (por ejemplo, la dirección del axón). Estos datos pueden llegar a seguir distribuciones de probabilidad muy complejas y pueden no ajustarse a ninguna distribución paramétrica conocida. El modelado de este tipo de problemas con redes bayesianas híbridas incluyendo variables discretas, lineales y direccionales presenta una serie de retos en relación al aprendizaje a partir de datos, la inferencia, etc. En esta tesis se propone un método para modelar y simular árboles dendríticos basales de neuronas piramidales usando redes bayesianas para capturar las interacciones entre las variables del problema. Para ello, se mide un amplio conjunto de variables de las dendritas y se aplica un algoritmo de aprendizaje con el que se aprende la estructura y se estiman los parámetros de las distribuciones de probabilidad que constituyen las redes bayesianas. Después, se usa un algoritmo de simulación para construir dendritas virtuales mediante el muestreo de valores de las redes bayesianas. Finalmente, se lleva a cabo una profunda evaluaci ón para verificar la capacidad del modelo a la hora de generar dendritas realistas. En esta primera aproximación, las variables fueron discretizadas para poder aprender y muestrear las redes bayesianas. A continuación, se aborda el problema del aprendizaje de redes bayesianas con diferentes tipos de variables. Las mixturas de polinomios constituyen un método para representar densidades de probabilidad en redes bayesianas híbridas. Presentamos un método para aprender aproximaciones de densidades unidimensionales, multidimensionales y condicionales a partir de datos utilizando mixturas de polinomios. El método se basa en interpolación con splines, que aproxima una densidad como una combinación lineal de splines. Los algoritmos propuestos se evalúan utilizando bases de datos artificiales. Además, las mixturas de polinomios son utilizadas como un método no paramétrico de estimación de densidades para clasificadores basados en redes bayesianas. Después, se estudia el problema de incluir información direccional en redes bayesianas. Este tipo de datos presenta una serie de características especiales que impiden el uso de las técnicas estadísticas clásicas. Por ello, para manejar este tipo de información se deben usar estadísticos y distribuciones de probabilidad específicos, como la distribución univariante von Mises y la distribución multivariante von Mises–Fisher. En concreto, en esta tesis extendemos el clasificador naive Bayes al caso en el que las distribuciones de probabilidad condicionada de las variables predictoras dada la clase siguen alguna de estas distribuciones. Se estudia el caso base, en el que sólo se utilizan variables direccionales, y el caso híbrido, en el que variables discretas, lineales y direccionales aparecen mezcladas. También se estudian los clasificadores desde un punto de vista teórico, derivando sus funciones de decisión y las superficies de decisión asociadas. El comportamiento de los clasificadores se ilustra utilizando bases de datos artificiales. Además, los clasificadores son evaluados empíricamente utilizando bases de datos reales. También se estudia el problema de la clasificación de interneuronas. Desarrollamos una aplicación web que permite a un grupo de expertos clasificar un conjunto de neuronas de acuerdo a sus características morfológicas más destacadas. Se utilizan medidas de concordancia para analizar el consenso entre los expertos a la hora de clasificar las neuronas. Se investiga la idoneidad de los términos anatómicos y de los tipos neuronales utilizados frecuentemente en la literatura a través del análisis de redes bayesianas y la aplicación de algoritmos de clustering. Además, se aplican técnicas de aprendizaje supervisado con el objetivo de clasificar de forma automática las interneuronas a partir de sus valores morfológicos. A continuación, se presenta una metodología para construir un modelo que captura las opiniones de todos los expertos. Primero, se genera una red bayesiana para cada experto y se propone un algoritmo para agrupar las redes bayesianas que se corresponden con expertos con comportamientos similares. Después, se induce una red bayesiana que modela la opinión de cada grupo de expertos. Por último, se construye una multired bayesiana que modela las opiniones del conjunto completo de expertos. El análisis del modelo consensuado permite identificar diferentes comportamientos entre los expertos a la hora de clasificar las neuronas. Además, permite extraer un conjunto de características morfológicas relevantes para cada uno de los tipos neuronales mediante inferencia con la multired bayesiana. Estos descubrimientos se utilizan para validar el modelo y constituyen información relevante acerca de la morfología neuronal. Por último, se estudia un problema de clasificación en el que la etiqueta de clase de los datos de entrenamiento es incierta. En cambio, disponemos de un conjunto de etiquetas para cada instancia. Este problema está inspirado en el problema de la clasificación de neuronas, en el que un grupo de expertos proporciona una etiqueta de clase para cada instancia de manera individual. Se propone un método para aprender redes bayesianas utilizando vectores de cuentas, que representan el número de expertos que seleccionan cada etiqueta de clase para cada instancia. Estas redes bayesianas se evalúan utilizando bases de datos artificiales de problemas de aprendizaje supervisado.

The Hidden benefits of “poster presentation assessment”: Skill acquirement and learning; evaluated from the nursing students perspective

Introduction
This paper reports to an exercise in evaluating poster group work and poster presentation and the extra learning and skill acquisition that this can provide to nursing students, through a creative and stimulating assessment method. Much had been written about the benefits of using posters as an assessment method, yet there appears to be a lack of research that captures the student experience.
Aim
This evaluative study sought to evaluate the student experience by using a triangulation approach to evaluation:
Methodology
All students from the February 2015 nursing intake, were eligible to take part (80 students) of which 71 participated (n=71). The poster group presentations took place at the end of their first phase of year one teaching and the evaluation took place at the end of their first year as undergraduate. Evaluation involved;
1. Quantitative data by questionnaires
2. Qualitative data from focus group discussions
Results
A number of key themes emerged from analysis of the data which captured the “added value” of learning from the process of poster assessment including:
 Professionalism: developing time keeping skills, presenting skills.
 Academic skills: developing literature search, critic and reporting
 Team building and collaboration
Overall 88% agreed that the process furnished them with additional skills and benefits above the actual production of the poster, with 97% agreeing that these additional skills are important skills for a nurse.
Conclusion
These results would suggest that the process of poster development and presentation furnish student nurses with many additional skills that they may not acquire through other types of assessment and are therefore beneficial. The structure of the assessment encourages a self-directed approach so students take control of the goals and purposes of learning. The sequential organization of the assessment guides students in the transition from dependent to self-directed learners.

Qualified intervention/qualifying formation

An education promoting scientific literacy (SL) that prepares the citizens to a responsible citizenship has persisted as an argument across discussions on curricula design. The ubiquity of science and technology on contemporary societies and the ideological requirement of informed democratic participation led to the identification of relevant categories that drive curriculum reforms towards a humanistic approach of school science. The category ‘Science as culture’ acquires in the current work a major importance: it enlightens the meaning of scientific literacy. Looking closely to the French term, culture scientifique et tecnologique, turns science simultaneously into a cultural object and product that can be both received and worked at different levels and within several approaches by the individuals and the communities. On the other hand, nonformal and informal education spaces gain greater importance. Together with the formal school environment these spaces allow for an enrichment and diversification of learning experiences. Examples of nonformal spaces where animators can develop their work may be science museums or botanical gardens; television and internet can be regarded as informal education spaces. Due to the above mentioned impossibility of setting apart the individual or community-based experiences from Science and Technology (S&T), the work in nonformal and informal spaces sets an additional challenge to the preparation of socio-cultural animators. Socio-scientific issues take, at times, heavily relevance within the communities. Pollution, high tension lines, spreading of diseases, food contamination or natural resources conservation are among the socio-scientific issues that often call upon arguments and emotions. In the context of qualifying programmes on socio-cultural animation (social education and community development) within European Higher Education Area (EHEA) the present study describes the Portuguese framework. The comparison of programmes within Portugal aims to contribute to the discussion on the curriculum design for a socio-cultural animator degree (1st cycle of Bologna process). In particular, this study intends to assess how the formation given complies with enabling animators to work, within multiple scenarios, with communities in situations of socio-scientific relevance. A set of themes, issues and both current and potential fields of action, not described or insufficiently described in literature, is identified and analysed in the perspective of a qualified intervention of animators. One of these examples is thoroughly discussed. Finally, suggestions are made about curriculum reforms in order, if possible, to strongly link the desired qualified intervention with a qualifying formation.

Challenges for e-learning and adult students in higher education

This article examines Lifelong Learning, from the perspective of the adult learner in higher education, by presenting some of the results of a project, funded by the European Commission's Socrates Programme, LIHE, Learning in Higher Education. It is structured as follows: first, the background of the project is described, then the experiences of the adult student, concerning their induction and tuition, are presented. Some future trends concerning adults in higher education and lifelong learning are outlined and conclusions drawn.

AzoresPhotoID: first results from an open-access photoidentification catalogue

25th Conference of the European Cetacean Society. Long-terms datasets on marine mammals: learning from the past to manage the future, Cadiz, Spain, 21-23 March 2011.

Ensemble: an e-learning framework

E-Learning frameworks are conceptual tools to organize networks of elearning services. Most frameworks cover areas that go beyond the scope of e-learning, from course to financial management, and neglects the typical activities in everyday life of teachers and students at schools such as the creation, delivery, resolution and evaluation of assignments. This paper presents the Ensemble framework - an e-learning framework exclusively focused on the teaching-learning process through the coordination of pedagogical services. The framework presents an abstract data, integration and evaluation model based on content and communications specifications. These specifications must base the implementation of networks in specialized domains with complex evaluations. In this paper we specialize the framework for two domains with complex evaluation: computer programming and computer-aided design (CAD). For each domain we highlight two Ensemble hotspots: data and evaluations procedures. In the former we formally describe the exercise and present possible extensions. In the latter, we describe the automatic evaluation procedures.

The development of a knowledge-based society - Challenges for e-learning and adult students in higher education

This chapter appears in Encyclopaedia of Distance Learning 2nd Edition edit by Rogers, P.; Berg, Gary; Boettecher, Judith V.; Howard, Caroline; Justice, Lorraine; Schenk, Karen D.. Copyright 2009, IGI Global, www.igi-global.com. Posted by permission of the publisher. URL: http://www.igi-global.com/reference/ details.asp?ID=9703&v=tableOfContents

From user browsing behaviour to user demographics

A Internet conta hoje com mais de 3 mil milhões de utilizadores e esse valor não para de aumentar. Desta forma, proporcionar uma experiência online agradável aos seus utilizadores é cada vez mais importante para as empresas. De modo a tirar partido dos benefícios deste crescimento, as empresas devem ser capazes de identificar os seus clientes-alvo dentro do total de utilizadores; e, subsequentemente, personalizar a sua experiência online. Existem diversas formas de estudar o comportamento online dos utilizadores; no entanto, estas não são ideais e existe uma ampla margem para melhoria. A inovação nesta área pode comportar um grande potencial comercial e até ser disruptiva. Com isto em mente, proponho-me a estudar a possível criacão de um sistema de aprendizagem automática (machine learning) que permita prever informa ações demográficas dos utilizadores estritamente com base no seu comportamento online. Tal sistema poderia constituir uma alternativa às atuais opções, que são mais invasivas; mitigando assim preocupações ao nível da proteção de dados pessoais. No primeiro capítulo (Introdução) explico a motivação para o estudo do comportamento dos utilizadores online por parte de empresas, e descrevo as opções disponíveis atualmente. Apresento também a minha proposta e o contexto em que assenta. O capítulo termina com a identicação de limitações que possam existir a priori. O segundo capítulo (Machine Learning) fornece uma introdução sobre machine learning, com o estudo dos algoritmos que vão ser utilizados e explicando como analisar os resultados. O terceiro capítulo (Implementação) explica a implementação do sistema proposto e descreve o sistema que desenvolvi no decorrer deste estudo, e como integra-lo em sistemas já existentes. No quarto capítulo (Análise e manipulação dos dados), mostro os dados compilados e explico como os recolhi e manipulei para testar a hipótese. No quinto capítulo (Análise de dados e discussão) vemos como e que os dados recolhidos foram usados pelos vários algoritmos para descobrir como se correlacionam com dados dos utilizadores e analiso e discuto os resultados observados. Por fim, o sexto e último capítulo apresenta as conclusões. Dependendo dos resultados, mostro como a hipótese poderia ser melhor testada, ou então discuto os próximos passos para tornar o sistema realidade.

O papel dos livros informativos para a infância na conformação de uma educação histórica

Relatório de estágio de mestrado em Educação Pré-Escolar e Ensino do 1ºCiclo do Ensino Básico

The normative dimension of social life from a second-person perspective

This thesis argues that insofar as we want to account for the normative dimension of social life, we must be careful to avoid construing that normative dimension in such a way as to exclude that which the second-person perspective reveals is important to social life and our ability to participate in it.¦The second-person perspective reveals that social life ought to be understood as a mix or balance of the regular and the irregular, where, in addition, those one interacts with are always to some extent experienced as other in a way that is neither immediately, nor perhaps ultimately, understandable. For persons to be able to participate in social life, conceived of in this way, they must have abilities that allow them to be, to some extent, hesitant and tentative in their relations with others, and thus tolerant of ambiguity, uncertainty and unpredictability, and responsive to and capable of learning from the otherness of others in the course of interacting with them.¦Incorporating the second-person perspective means we have to make some changes to the way we think about the normative in general, and the normative dimension of social life in particular. It does not mean giving up on the distinction between the normative and the regular - that continues to be fundamentally important but it does mean not excluding, as part of social life and as worthy of explanation, all that which is irregular. A radical way of putting it would be to say that there must be a sense in which the irregular is part of the normative. A less radical way, and the way adopted by this thesis, is to say that any account of the normative dimension of social life must not be such as to exclude the importance of irregularity from social life. This will mean 1) not characterising conventions, norms and rules as determinants of appropriateness and inappropriateness; 2) not thinking of them as necessary; 3) not thinking of them as necessarily governing minds; and 4) not thinking of them as necessarily shared.¦-¦L'argument principal de la thèse est que, pour rendre compte de la dimension normative de la vie sociale, il faut veiller à ne pas exclure la perspective de la deuxième personne - une perspective importante pour comprendre la vie sociale et la capacité requise pour y participer.¦Cette perspective nous permet d'imaginer la vie sociale comme un mélange ou un équilibre entre le régulier et l'irrégulier, l'interaction entre des individus pouvant être appréhendée comme l'expérience de chaque personne avec «l'autre» d'une manière qui n'est pas immédiatement compréhensible, et qui ne peut pas, peut-être, être ultimement comprise. Pour participer à la vie sociale, l'on doit avoir la capacité de rester hésitant et «réactif» dans ses relations avec les autres, de rester ouvert à leur altérité et de tolérer l'ambiguïté, l'incertitude et l'imprévisibilité des interactions sociales.¦Adopter une perspective «à la deuxième personne» conduit à une autre manière de penser la normativité en général, et la dimension normative de la vie sociale en particulier. Cela ne veut pas dire qu'il faut abandonner la distinction entre le normatif et le régulier - une distinction qui garde une importance fondamentale - mais qu'il faut reconnaître l'irrégulier comme faisant partie de la vie sociale et comme étant digne, en tant que tel, d'être expliqué. Une conception radicale pourrait même concevoir l'irrégulier comme faisant partie intégrante de la normativité. Une approche moins radicale, qui est celle adoptée dans cette thèse, est de dire que tout compte-rendu de la dimension normative de la vie sociale doit prendre en considération l'importance de l'irrégularité dans la vie sociale. Une telle approche implique que les conventions, normes et règles (1) ne déterminent pas ce qui est approprié ou inapproprié; (2) ne sont pas toujours nécessaires ; (3) ne gouvernent pas le fonctionnement de l'esprit ; et (4) ne sont pas nécessairement partagées.