Search

Animal affective computing is a quickly growing field of research, where only recently first efforts to go beyond animal tracking into recognizing their internal states, such as pain and emotions, have emerged. In most mammals, facial expressions are an important channel for communicating information about these states. However, unlike the human domain, there is an acute lack of datasets that make automation of facial analysis of animals feasible. This paper aims to fill this gap by presenting a dataset called Cat Facial Landmarks in the Wild (CatFLW) which contains 2016 images of cat faces in different environments and conditions, annotated with 48 facial landmarks specifically chosen for their relationship with underlying musculature, and relevance to cat-specific facial Action Units (CatFACS). To the best of our knowledge, this dataset has the largest amount of cat facial landmarks available. In addition, we describe a semi-supervised (human-in-the-loop) method of annotating images with landmarks, used for creating this dataset, which significantly reduces the annotation time and could be used for creating similar datasets for other animals. The dataset is available on request.

Internal affective states are closely linked to facial expressions in both human and many non-human animals. For some animal species, objective tools for facial expression analysis such as AnimalFACS are available and are only just beginning to be increasingly used. However, their use requires special expertise, training, and certification, while there is still some remaining risk for human bias, and its application is time-consuming. Automation of facial analysis offers a promising alternative and is already being addressed by a large body of research in the human domain. In animal research, automation has so far been addressed for a few species in the context of pain, while emotional state recognition remains underexplored, especially in canine species due to the complexity of their facial morphology and expressions. The contribution of the present study is twofold. First, this is the first study to address automated recognition of emotional states in dogs using a dataset obtained in a controlled experimental setting, including videos from (n=29) Labrador Retrievers assumed to be in two experimentally induced emotional states: negative (frustration) and positive (anticipation). Two different approaches are compared in relation to our aim: (1) A DogFACS-based approach with a two-step pipeline consisting of (i) an Action Unit (AU) detector and (ii) a positive/negative state Decision Tree classifier; (2) An approach using deep learning techniques with no intermediate representation. The approaches reach accuracy of above 71% and 89%, respectively, with the deep learning approach performing better. Secondly, this study is the first to address the explainability of AI models in the context of understanding the expression of emotion in animals. The DogFACS-based approach provides decision trees, that is a mathematical representation which reflects previous findings by human experts in relation to certain DogFACS variables being correlates of specific emotional states. The deep learning approach offers a different, visual form of explainability in the form of heatmaps reflecting regions of focus of the network's attention, which in some cases show focus clearly related to the nature of particular AUs. These heatmaps may hold the key to novel insights on the sensitivity of the network to nuanced pixel patterns reflecting information invisible to the human eye.

Similarly to humans, facial expressions in animals are closely linked with emotional states. However, in contrast to the human domain, automated recognition of emotional states from facial expressions in animals is underexplored, mainly due to difficulties in data collection and establishment of ground truth concerning emotional states of non-verbal users. We apply recent deep learning techniques to classify (positive) anticipation and (negative) frustration of dogs on a dataset collected in a controlled experimental setting. We explore the suitability of different backbones (e.g. ResNet, ViT) under different supervisions to this task, and find that features of a self-supervised pretrained ViT (DINO-ViT) are superior to the other alternatives. To the best of our knowledge, this work is the first to address the task of automatic classification of canine emotions on data acquired in a controlled experiment.

Facial expressions in non-human animals are closely linked to their internal affective states , with the majority of empirical work focusing on facial shape changes associated with pain. However, existing tools for facial expression analysis are prone to human subjectivity and bias, and in many cases also require special expertise and training. This paper presents the first comparative study of two different paths towards automatizing pain recognition in facial images of domestic short haired cats (n = 29), captured during ovariohysterectomy at different time points corresponding to varying intensities of pain. One approach is based on convolutional neural networks (ResNet50), while the other – on machine learning models based on geometric landmarks analysis inspired by species specific Facial Action Coding Systems (i.e. catFACS). Both types of approaches reach comparable accuracy of above 72%, indicating their potential usefulness as a basis for automating cat pain detection from images.

Advances in animal motion tracking and pose recognition have been a game changer in the study of animal behavior. Recently, an increasing number of works go 'deeper' than tracking, and address automated recognition of animals' internal states such as emotions and pain with the aim of improving animal welfare, making this a timely moment for a systematization of the field. This paper provides a comprehensive survey of computer vision-based research on recognition of affective states and pain in animals, addressing both facial and bodily behavior analysis. We summarize the efforts that have been presented so far within this topic -- classifying them across different dimensions, highlight challenges and research gaps, and provide best practice recommendations for advancing the field, and some future directions for research.

Animal-assisted reading is a form of animal-assisted activity where children interact with and read aloud to specially trained therapy dogs. Such activities have been shown to have positive impact on literacy skills, reading motivation and sense of empowerment and self-esteem in children. Mobile technologies and facets of engagement in digital reading are increasingly studied in the context of child learning and education. This short paper presents some first results of our ongoing study exploring potential ways in which mobile technology can be used to enhance and support animal-assisted reading activities for children.

This research-in-progress paper proposes the need for a move towards more meaningful variability of visual notations. Evidence accumulated via an online survey on the requirements practitioners have for visual notations, indicate the need for variability of a modeling language's visual notation. Widely used modeling languages in practice such as UML and BPMN do not support redesign of the visual notation of core constructs without modifying or extending the underlying abstract syntax and semantics (e.g., UML stereotypes, BPMN extensions). The expressed need to vary commonly used visual notations depending on particular users or contexts, while not changing the underlying modeling language itself, poses a set of research challenges discussed here.

The Physics of Notations [9] (PoN) is a design theory presenting nine principles that can be used to evaluate and improve the cognitive effectiveness of a visual notation. The PoN has been used to analyze existing standard visual notations (such as BPMN, UML, etc.), and is commonly used for evaluating newly introduced visual notations and their extensions. However, due to the rather vague and abstract formulation of the PoN's principles, they have received different interpretations in their operationalization. To address this problem, there have been attempts to formalize the principles, however only a very limited number of principles was covered. This research-in-progress paper aims to better understand the difficulties inherent in operationalizing the PoN, and better separate aspects of PoN, which can potentially be formulated in mathematical terms from those grounded in user-specific considerations.

In contrast to Computer Science, where the fundamental role of Logic is widely recognized, it plays a practically non-existent role in Information Systems curricula. In this paper we argue that instead of Logic's exclusion from the IS curriculum, a significant adaptation of the contents, as well as teaching methodologies, is required for an alignment with the needs of IS practitioners. We present our vision for such adaptation and report on concrete steps towards its implementation in the design and teaching of a course for graduate IS students at the University of Haifa. We discuss the course plan and present some data on the students' feedback on the course., Proceedings of the Fourth International Conference on Tools for Teaching Logic (TTL2015), Rennes, France, June 9-12, 2015. Editors: M. Antonia Huertas, Jo\~ao Marcos, Mar\'ia Manzano, Sophie Pinchinat, Fran\c{c}ois Schwarzentruber

Orientador: Marcelo Esteban Coniglio Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Filosofia e Ciências Humanas Resumo: Este trabalho está dividido entre duas grandes áreas: na teoria de multiálgebras, focamos majoritariamente em uma nova definição do que um objeto livremente gerado deveria ser em sua categoria e em como esta categoria é equivalente a outra com álgebras parcialmente ordenadas como objetos; então usamos semânticas não-determinísticas, especialmente aquela que nomeamos Nmatrizes restritas, nas lógicas paraconsistentes e em alguns sistemas lidando com uma nova apresentação do conceito natural de incompatibilidade, que generaliza o conceito de inconsistência. Em álgebra, nos focaremos nas não-determinísticas, também conhecidas como multiálgebras, cujas operações retornam subconjuntos não vazios de seus universos. Enquanto a categoria de álgebras sobre uma assinatura possui objetos livremente gerados, os quais permitem em certo sentido a extensão única de funções a homomorfismos, a categoria de multiálgebras sobre uma assinatura dada não possui elementos com propriedades comparáveis. Para contornar este problema, estendemos o significado de uma álgebra de fórmulas: se uma multiálgebra generaliza uma álgebra ao ter múltiplos resultados para uma dada operação, uma multiálgebra de fórmulas generaliza uma álgebra de fórmulas ao ter múltiplas possibilidade para a aplicação de um conectivo a fórmulas dadas. Quanto à categoria de multiálgebras propriamente dita, oferecemos uma equivalência entre ela e uma categoria livre de não-determinismo, que alternativamente possui álgebras ordenadas, semelhantes a álgebras de Boole, como objetos. Na parte dedicada à lógica, nossos objetivos são novamente dois: primeiramente, algumas lógicas de inconsistência formal, exempli gratia aquelas da hierarquia de da Costa, não podem ser caracterizadas por Nmatrizes finitas. No que é um desenvolvimento muito natural, uma Nmatriz restrita, ou RNmatriz, restringe aquelas homomorfismos que devem ser considerados quando testamos a validade de uma dedução por uma Nmatriz. Mostramos como esta distinção provê as RNmatrizes finitas com poder expressivo muito superior, suficientente para adequadamente caracterizar os sistemas de da Costa e dar a eles métodos de decisão, tanto baseados em tabelas de verdade quanto em semânticas de tableaux. Em outra direção, generalizamos as lógicas de inconsistência formal a sistemas construídos em torno da noção de incompatibilidade: o Leitmotiv sendo que duas fórmulas incompatíveis simultaneamente verdadeiras trivializam uma dedução, e como um caso especial, uma fórmula é consistente quando é incompatível com sua negação. Mostramos como essa noção estende aquela de inconsistência de maneira não-trivial, apresentando traduções conservativas para muitos dos sistemas inconsistentes mais simples em lógicas de incompatilidade, apresentamos semânticas construídas com RNmatrizes para essas novas lógicas e mostramos que elas não podem ser caracterizadas por métodos mais usuais Abstract: This work is divided between two main areas: in the theory of multialgebras, we focus mostly on a new definition of what a freely generated object should be in their category, and on how this category is equivalent to another with partially ordered algebras as objects; we then use nondeterministic semantics, specially those we have named restricted Nmatrices, on paraconsistent logics and some systems dealing with a new presentation of the natural concept of incompatibility, which generalizes inconsistency. In algebra, we will focus on the non-deterministic ones, also known as multialgebras, whose operations return non-empty subsets of their universes. While the category of algebras over a signature has freely generated objects, which in a sense permit for the unique extension of functions to homomorphisms, the category of multialgebras over a given signature does not have elements with comparable properties. To circumvent this problem, we widen our understanding of algebras of formulas: if a multialgebra generalizes an algebra by having multiple results for a given operation, a multialgebra of formulas should generalize an algebra of formulas by having multiple possibilities for applying a connective to given formulas. Concerning the category of multialgebras itself, we offer an equivalence between it and a category avoiding non-determinism altogether, relying instead on ordered Boolean-like algebras as objects. On the part devoted to logic, our goals are again roughly twofold: firstly, some logics of formal inconsistency, exempli gratia those found in da Costa’s hierarchy, cannot be characterized by finite Nmatrices. In what is a very natural development, a restricted Nmatrix (or RNmatrix) restricts those homomorphisms to be taken into consideration when evaluating the validity of a deduction according to an Nmatrix. We show how this distinction gives far greater expressiveness to finite RNmatrices, enough to adequately characterize da Costa’s systems and provide decision methods for those logics, both based on row-branching, row-eliminating truth tables, and tableau semantics. In another direction, we generalize logics of formal inconsistency to new systems built around the notion of incompatibility: the Leitmotiv being that having two incompatible formulas to simultaneously hold trivializes a deduction, and as a special case, a formula is consistent when it is incompatible with its negation. We show how this notion extends that of inconsistency in a non-trivial way, presenting conservative translations for many simple inconsistent systems into logics of incompatibility; we also provide semantics built on RNmatrices for these new logics, and prove that they can not be characterized by more standard methods Doutorado Filosofia Doutor em Filosofia CAPES 88882.329848/2019