Event detection is a critical challenge in many fields like video surveillance, social graph analysis, and multimedia processing. Furthermore, events are “structured” objects involv ing multiple components like the event type, the participants with their roles, and the atomic events in which it decomposes. Therefore, the recognition of an event is not only limited to recognize the type of the event and when it happened, but it involves solving a set of simple tasks. Exploiting background knowledge about events and their relations could then be beneficial for event detection. In the last years, neuro-symbolic integration has been proposed to merge the strengths and overcome the drawbacks of both symbolic and neural worlds. As a consequence, different neuro-symbolic frameworks, which com bine low-level perception of neural networks with a symbolic layer, encoding prior domain knowledge (usually defined in terms of logical rules), have been applied to solve different atemporal tasks. In this thesis, we want to investigate the application of the neuro-symbolic paradigm for event detection. This would also provide a better insight into the strengths and limitations of neuro-symbolic towards tasks involving time.

A inteligência computacional é considerada por diferentes autores da atualidade como o destino manifesto da Ciência da Computação. A modelagem de diversos aspectos da cognição, tais como aprendizagem e raciocínio, tem sido a motivação para o desenvolvimento dos paradigmas simbólico e conexionista da inteligência artificial e, mais recentemente, para a integração de ambos com o intuito de unificar as vantagens de cada abordagem em um modelo único. Para o desenvolvimento de sistemas inteligentes, bem como para diversas outras áreas da Ciência da Computação, o tempo é considerado como um componente essencial, e a integração de uma dimensão temporal nestes sistemas é fundamental para conseguir uma representação melhor do comportamento cognitivo. Neste trabalho, propomos o SCTL (Sequential Connectionist Temporal Logic), uma abordagem neuro-simbólica para integrar conhecimento temporal, representado na forma de programas em lógica, em redes neurais recorrentes, de forma que a caracterização semântica de ambas representações sejam equivalentes. Além da estratégia para realizar esta conversão entre representações, e da verificação formal da equivalência semântica, também realizamos uma comparação da estratégia proposta com relação a outros sistemas que realizam representação simbólica e temporal em redes neurais. Por outro lado, também descrevemos, de foma algorítmica, o comportamento desejado para as redes neurais geradas, para realizar tanto inferência quanto aprendizagem sob uma ótica temporal. Este comportamento é analisado em diversos experimentos, buscando comprovar o desempenho de nossa abordagem para a modelagem cognitiva considerando diferentes condições e aplicações. / Computational Intelligence is considered, by di erent authors in present days, the manifest destiny of Computer Science. The modelling of di erent aspects of cognition, such as learning and reasoning, has been a motivation for the integrated development of the symbolic and connectionist paradigms of artificial intelligence. More recently, such integration has led to the construction of models catering for integrated learning and reasoning. The integration of a temporal dimension into such systems is a relevant task as it allows for a richer representation of cognitive behaviour features, since time is considered an essential component in intelligent systems development. This work introduces SCTL (Sequential Connectionist Temporal Logic), a neuralsymbolic approach for integrating temporal knowledge, represented as logic programs, into recurrent neural networks. This integration is done in such a way that the semantic characterization of both representations are equivalent. Besides the strategy to achieve translation from one representation to another, and verification of the semantic equivalence, we also compare the proposed approach to other systems that perform symbolic and temporal representation in neural networks. Moreover, we describe the intended behaviour of the generated neural networks, for both temporal inference and learning through an algorithmic approach. Such behaviour is then evaluated by means several experiments, in order to analyse the performance of the model in cognitive modelling under di erent conditions and applications.

A inteligência computacional é considerada por diferentes autores da atualidade como o destino manifesto da Ciência da Computação. A modelagem de diversos aspectos da cognição, tais como aprendizagem e raciocínio, tem sido a motivação para o desenvolvimento dos paradigmas simbólico e conexionista da inteligência artificial e, mais recentemente, para a integração de ambos com o intuito de unificar as vantagens de cada abordagem em um modelo único. Para o desenvolvimento de sistemas inteligentes, bem como para diversas outras áreas da Ciência da Computação, o tempo é considerado como um componente essencial, e a integração de uma dimensão temporal nestes sistemas é fundamental para conseguir uma representação melhor do comportamento cognitivo. Neste trabalho, propomos o SCTL (Sequential Connectionist Temporal Logic), uma abordagem neuro-simbólica para integrar conhecimento temporal, representado na forma de programas em lógica, em redes neurais recorrentes, de forma que a caracterização semântica de ambas representações sejam equivalentes. Além da estratégia para realizar esta conversão entre representações, e da verificação formal da equivalência semântica, também realizamos uma comparação da estratégia proposta com relação a outros sistemas que realizam representação simbólica e temporal em redes neurais. Por outro lado, também descrevemos, de foma algorítmica, o comportamento desejado para as redes neurais geradas, para realizar tanto inferência quanto aprendizagem sob uma ótica temporal. Este comportamento é analisado em diversos experimentos, buscando comprovar o desempenho de nossa abordagem para a modelagem cognitiva considerando diferentes condições e aplicações. / Computational Intelligence is considered, by di erent authors in present days, the manifest destiny of Computer Science. The modelling of di erent aspects of cognition, such as learning and reasoning, has been a motivation for the integrated development of the symbolic and connectionist paradigms of artificial intelligence. More recently, such integration has led to the construction of models catering for integrated learning and reasoning. The integration of a temporal dimension into such systems is a relevant task as it allows for a richer representation of cognitive behaviour features, since time is considered an essential component in intelligent systems development. This work introduces SCTL (Sequential Connectionist Temporal Logic), a neuralsymbolic approach for integrating temporal knowledge, represented as logic programs, into recurrent neural networks. This integration is done in such a way that the semantic characterization of both representations are equivalent. Besides the strategy to achieve translation from one representation to another, and verification of the semantic equivalence, we also compare the proposed approach to other systems that perform symbolic and temporal representation in neural networks. Moreover, we describe the intended behaviour of the generated neural networks, for both temporal inference and learning through an algorithmic approach. Such behaviour is then evaluated by means several experiments, in order to analyse the performance of the model in cognitive modelling under di erent conditions and applications.

A inteligência computacional é considerada por diferentes autores da atualidade como o destino manifesto da Ciência da Computação. A modelagem de diversos aspectos da cognição, tais como aprendizagem e raciocínio, tem sido a motivação para o desenvolvimento dos paradigmas simbólico e conexionista da inteligência artificial e, mais recentemente, para a integração de ambos com o intuito de unificar as vantagens de cada abordagem em um modelo único. Para o desenvolvimento de sistemas inteligentes, bem como para diversas outras áreas da Ciência da Computação, o tempo é considerado como um componente essencial, e a integração de uma dimensão temporal nestes sistemas é fundamental para conseguir uma representação melhor do comportamento cognitivo. Neste trabalho, propomos o SCTL (Sequential Connectionist Temporal Logic), uma abordagem neuro-simbólica para integrar conhecimento temporal, representado na forma de programas em lógica, em redes neurais recorrentes, de forma que a caracterização semântica de ambas representações sejam equivalentes. Além da estratégia para realizar esta conversão entre representações, e da verificação formal da equivalência semântica, também realizamos uma comparação da estratégia proposta com relação a outros sistemas que realizam representação simbólica e temporal em redes neurais. Por outro lado, também descrevemos, de foma algorítmica, o comportamento desejado para as redes neurais geradas, para realizar tanto inferência quanto aprendizagem sob uma ótica temporal. Este comportamento é analisado em diversos experimentos, buscando comprovar o desempenho de nossa abordagem para a modelagem cognitiva considerando diferentes condições e aplicações. / Computational Intelligence is considered, by di erent authors in present days, the manifest destiny of Computer Science. The modelling of di erent aspects of cognition, such as learning and reasoning, has been a motivation for the integrated development of the symbolic and connectionist paradigms of artificial intelligence. More recently, such integration has led to the construction of models catering for integrated learning and reasoning. The integration of a temporal dimension into such systems is a relevant task as it allows for a richer representation of cognitive behaviour features, since time is considered an essential component in intelligent systems development. This work introduces SCTL (Sequential Connectionist Temporal Logic), a neuralsymbolic approach for integrating temporal knowledge, represented as logic programs, into recurrent neural networks. This integration is done in such a way that the semantic characterization of both representations are equivalent. Besides the strategy to achieve translation from one representation to another, and verification of the semantic equivalence, we also compare the proposed approach to other systems that perform symbolic and temporal representation in neural networks. Moreover, we describe the intended behaviour of the generated neural networks, for both temporal inference and learning through an algorithmic approach. Such behaviour is then evaluated by means several experiments, in order to analyse the performance of the model in cognitive modelling under di erent conditions and applications.

In this thesis, we discuss different techniques to bridge the gap between two different approaches to artificial intelligence: the symbolic and the connectionist paradigm. Both approaches have quite contrasting advantages and disadvantages. Research in the area of neural-symbolic integration aims at bridging the gap between them. Starting from a human readable logic program, we construct connectionist systems, which behave equivalently. Afterwards, those systems can be trained, and later the refined knowledge be extracted.

Une étude de terrain a été accomplie dans le milieu associatif immigrant féminin de Montréal afin d’investiguer le rôle que peut avoir la participation à une association de femmes immigrantes quant à l’intégration de celles-ci à leur nouvelle société. Deux associations ont été ciblées pour cette étude : le Centre Femmes du monde à Côte-des-Neiges et le Comité des femmes des communautés culturelles, issu de la Fédération des femmes du Québec. Le premier est un organisme communautaire de quartier et le second, un groupe de défense et de revendication de droits des femmes immigrantes, à l’échelle de la province. Une période d’observation participante s’échelonnant de février 2007 à juin 2008 ainsi que 21 entrevues individuelles auprès de participantes ont été réalisées. L’analyse de ces données montre que la participation contribue, d’une manière tantôt similaire, tantôt distincte à l’intérieur des deux espaces de participation, à différentes dimensions de l’intégration des participantes : l’adaptation fonctionnelle, l’intégration sociale et plus particulièrement l’intégration symbolique. L’aspect symbolique de l’intégration, discuté en profondeur dans ce mémoire, sous-tend les idées de développement d’un sentiment d’appartenance et de reconnaissance sociale à la fois individuelle et collective des femmes immigrantes à l’intérieur de leur nouvelle société. / Fieldwork was carried out in immigrant women’s associations in Montreal to investigate the role of participation of immigrant women in such associations for their integration to their new society. Two associations have been targeted for this study: a neighborhood community association, the Centre Femmes du monde à Côte-des-Neiges, and the Comité des femmes des communautés culturelles of the Fédération des femmes du Québec, a group that defends immigrant women’s rights, at the provincial level. Participant observation was done between February 2007 and June 2008, and a series of 21 individual interviews were completed. Analysis of the data shows that participation in both associations contributes, in similar yet distinct ways, to various aspects of the participants’ integration to the host society: functional adaptation, social integration and, in particular, symbolic integration. This symbolic aspect of integration, which is extensively discussed throughout the thesis, underlies the development of a feeling of belonging and of individual and collective social recognition of immigrant women in their new society.

Wir untersuchen Feynman-Integrale in der Darstellung mit Schwinger-Parametern und leiten rekursive Integralgleichungen für masselose 3- und 4-Punkt-Funktionen her. Eigenschaften der analytischen (und dimensionalen) Regularisierung werden zusammengefasst und wir beweisen, dass in der Euklidischen Region jedes Feynman-Integral als eine Linearkombination konvergenter Feynman-Integrale geschrieben werden kann. Dies impliziert, dass man stets eine Basis aus konvergenten Masterintegralen wählen kann und somit divergente Integrale nicht selbst berechnet werden müssen. Weiterhin geben wir eine in sich geschlossene Darstellung der Theorie der Hyperlogarithmen und erklären detailliert die nötigen Algorithmen, um diese für die Berechnung mehrfacher Integrale anzuwenden. Wir definieren eine neue Methode um die Singularitäten solcher Integrale zu bestimmen und stellen ein Computerprogramm vor, welches die Integrationsalgorithmen implementiert. Unser Hauptresultat ist die Konstruktion unendlicher Familien masseloser 3- und 4-Punkt-Funktionen (diese umfassen unter anderem alle Leiter-Box-Graphen und deren Minoren), deren Feynman-Integrale zu allen Ordnungen in der epsilon-Entwicklung durch multiple Polylogarithmen dargestellt werden können. Diese Integrale können mit dem vorgestellten Programm explizit berechnet werden. Die Arbeit enthält interessante Beispiele von expliziten Ergebnissen für Feynman-Integrale mit bis zu 6 Schleifen. Insbesondere präsentieren wir den ersten exakt bestimmten Gegenterm in masseloser phi^4-Theorie, der kein multipler Zetawert ist sondern eine Linearkombination multipler Polylogarithmen, ausgewertet an primitiven sechsten Einheitswurzeln (und geteilt durch die Quadratwurzel aus 3). Zu diesem Zweck beweisen wir ein Paritätsresultat über die Zerlegbarkeit der Real- und Imaginärteile solcher Zahlen in Produkte und Beiträge geringerer Tiefe (depth). / We study Feynman integrals in the representation with Schwinger parameters and derive recursive integral formulas for massless 3- and 4-point functions. Properties of analytic (including dimensional) regularization are summarized and we prove that in the Euclidean region, each Feynman integral can be written as a linear combination of convergent Feynman integrals. This means that one can choose a basis of convergent master integrals and need not evaluate any divergent Feynman graph directly. Secondly we give a self-contained account of hyperlogarithms and explain in detail the algorithms needed for their application to the evaluation of multivariate integrals. We define a new method to track singularities of such integrals and present a computer program that implements the integration method. As our main result, we prove the existence of infinite families of massless 3- and 4-point graphs (including the ladder box graphs with arbitrary loop number and their minors) whose Feynman integrals can be expressed in terms of multiple polylogarithms, to all orders in the epsilon-expansion. These integrals can be computed effectively with the presented program. We include interesting examples of explicit results for Feynman integrals with up to 6 loops. In particular we present the first exactly computed counterterm in massless phi^4 theory which is not a multiple zeta value, but a linear combination of multiple polylogarithms at primitive sixth roots of unity (and divided by the square-root of 3). To this end we derive a parity result on the reducibility of the real- and imaginary parts of such numbers into products and terms of lower depth.

Une étude de terrain a été accomplie dans le milieu associatif immigrant féminin de Montréal afin d’investiguer le rôle que peut avoir la participation à une association de femmes immigrantes quant à l’intégration de celles-ci à leur nouvelle société. Deux associations ont été ciblées pour cette étude : le Centre Femmes du monde à Côte-des-Neiges et le Comité des femmes des communautés culturelles, issu de la Fédération des femmes du Québec. Le premier est un organisme communautaire de quartier et le second, un groupe de défense et de revendication de droits des femmes immigrantes, à l’échelle de la province. Une période d’observation participante s’échelonnant de février 2007 à juin 2008 ainsi que 21 entrevues individuelles auprès de participantes ont été réalisées. L’analyse de ces données montre que la participation contribue, d’une manière tantôt similaire, tantôt distincte à l’intérieur des deux espaces de participation, à différentes dimensions de l’intégration des participantes : l’adaptation fonctionnelle, l’intégration sociale et plus particulièrement l’intégration symbolique. L’aspect symbolique de l’intégration, discuté en profondeur dans ce mémoire, sous-tend les idées de développement d’un sentiment d’appartenance et de reconnaissance sociale à la fois individuelle et collective des femmes immigrantes à l’intérieur de leur nouvelle société. / Fieldwork was carried out in immigrant women’s associations in Montreal to investigate the role of participation of immigrant women in such associations for their integration to their new society. Two associations have been targeted for this study: a neighborhood community association, the Centre Femmes du monde à Côte-des-Neiges, and the Comité des femmes des communautés culturelles of the Fédération des femmes du Québec, a group that defends immigrant women’s rights, at the provincial level. Participant observation was done between February 2007 and June 2008, and a series of 21 individual interviews were completed. Analysis of the data shows that participation in both associations contributes, in similar yet distinct ways, to various aspects of the participants’ integration to the host society: functional adaptation, social integration and, in particular, symbolic integration. This symbolic aspect of integration, which is extensively discussed throughout the thesis, underlies the development of a feeling of belonging and of individual and collective social recognition of immigrant women in their new society.

Artificial neural networks (ANNs) and logic programs have both been suggested as means of modelling human cognition. While ANNs are adaptable and relatively noise resistant, the information they represent is distributed across various neurons and is therefore difficult to interpret. On the contrary, symbolic systems such as logic programs are interpretable but less adaptable. Human cognition is performed in a network of biological neurons and yet is capable of representing symbols, and therefore an ideal model would combine the strengths of the two approaches. This is the goal of Neural-Symbolic Integration [4, 16, 21, 40], in which ANNs are used to produce interpretable, adaptable representations of logic programs and other symbolic models. One neural-symbolic model of reasoning is SHRUTI [89, 95], argued to exhibit biological plausibility in that it captures some aspects of real biological processes. SHRUTI's original developers also suggest that further biological plausibility can be ascribed to the fact that SHRUTI networks can be represented by a model of genetic development [96, 120]. The aims of this thesis are to support the claims of SHRUTI's developers by producing the first such genetic representation for SHRUTI networks and to explore biological plausibility further by investigating the evolvability of the proposed SHRUTI genome. The SHRUTI genome is developed and evolved using principles from Generative and Developmental Systems and Artificial Development [13, 105], in which genomes use indirect encoding to provide a set of instructions for the gradual development of the phenotype just as DNA does for biological organisms. This thesis presents genomes that develop SHRUTI representations of logical relations and episodic facts so that they are able to correctly answer questions on the knowledge they represent. The evolvability of the SHRUTI genomes is limited in that an evolutionary search was able to discover genomes for simple relational structures that did not include conjunction, but could not discover structures that enabled conjunctive relations or episodic facts to be learned. Experiments were performed to understand the SHRUTI fitness landscape and demonstrated that this landscape is unsuitable for navigation using an evolutionary search. Complex SHRUTI structures require that necessary substructures must be discovered in unison and not individually in order to yield a positive change in objective fitness that informs the evolutionary search of their discovery. The requirement for multiple substructures to be in place before fitness can be improved is probably owed to the localist representation of concepts and relations in SHRUTI. Therefore this thesis concludes by making a case for switching to more distributed representations as a possible means of improving evolvability in the future.

In this thesis, we discuss different techniques to bridge the gap between two different approaches to artificial intelligence: the symbolic and the connectionist paradigm. Both approaches have quite contrasting advantages and disadvantages. Research in the area of neural-symbolic integration aims at bridging the gap between them. Starting from a human readable logic program, we construct connectionist systems, which behave equivalently. Afterwards, those systems can be trained, and later the refined knowledge be extracted.