Notre projet vise à suivre et reconstruire, de l’intérieur de la pensée de Merleau-Ponty, le concept de « personnel » dans sa relation avec ceux d’« impersonnel » et de « généralité », les trois étant impliqués dans la notion de « chair » et faisant partie de la reformulation du sujet et de l’identité. Selon nous, négliger la connexion existante entre les aspects les plus « personnels » de chaque individu et ceux plus « généraux », qui soutiennent une conception aussi bien du sujet que de la société, donne lieu à une lecture partielle de l’auteur. Cela risque, de surcroît, de rendre difficile la compréhension de la dimension qui en est la synthèse la plus avancée chez Merleau-Ponty, à savoir sa dimension « charnelle ». Étudier la relation qui entrelace le « personnel » au « général », en passant par l’analyse du concept d’« anonymat », revient à chercher une clef d’intelligibilité majeure de la pensée merleau-pontienne, qui respecte l’ampleur de sa réflexion, pour en entrevoir les raisons de départ et en justifier les points d’arrivée, grâce à l’approfondissement de ses textes encore inédits. / Our project aims to reconstitute, from the inside of Merleau-Ponty’s philosophical thought, the concept of « personal » in relation to the concepts of « impersonal » and « generality ». Those three concepts are involved in the notion of « flesh » and are part of the reformulation of the terms of subject (person) and identity. According to us, neglecting the bond between the more « personal » aspects of an individual and those more « general », aspects which support a conception of both subject and society, would lead to a « partial » interpretation of our author. In addition, it would make difficult the understanding of Merleau-Ponty’s most advanced synthesis, namely the dimension of the flesh. Studying the relation that intertwines the « personal » with the « general », through the examination of the importance of the notion of « anonymity » for our author, would mean instead looking for a central key of reading in the merleau-pontian thought, that would respect the whole amplitude of its reflection, so to sense the starting reasons and justify the arriving points, thanks to the in-depth study of its still unpublished texts.

Submitted by Ana Caroline Costa (ana_caroline212@hotmail.com) on 2018-11-13T19:24:29Z No. of bitstreams: 2 Dissertação - Bruna Garcia da Silveira Miguel Elias - 2018.pdf: 1464993 bytes, checksum: 4fb42bf2c5bf3dba934d5546c6206189 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-11-14T12:53:24Z (GMT) No. of bitstreams: 2 Dissertação - Bruna Garcia da Silveira Miguel Elias - 2018.pdf: 1464993 bytes, checksum: 4fb42bf2c5bf3dba934d5546c6206189 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-11-14T12:53:24Z (GMT). No. of bitstreams: 2 Dissertação - Bruna Garcia da Silveira Miguel Elias - 2018.pdf: 1464993 bytes, checksum: 4fb42bf2c5bf3dba934d5546c6206189 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-10-15 / This dissertation is an investigation of the notions of “indeterminacy”, “inexactness” and “vagueness” in the context of Wittgenstein's phenomenology during in the intermediary period of the author. The occurrence of vague terms in phenomenological expressions seems a problem to be solved, since it gives rise to some dissatisfaction with the lack of “precision” which should be a feature of any suitable language. We will deal with this “problem” about the vagueness expressions and some reflections about it, specifically as they were presented by Wittgenstein in the work Phenomenal Language. We explore in this dissertation some changes in the thought of this philosopher, from the Tractatus-Logico-Philosophicus to the text Phenomenal Language, with the intention of identifying, clarifying and later proposing modes to overcome the mistakes from the new ideas of Wittgenstein throughout the intermediate period, as they are rejected by the philosopher himself in the mentioned text. Such illusions, according to Wittgenstein, would be responsible for certain illusory conceptions about “language” and “perception” itself. We shall deal with some of Wittgenstein's notes on these misleading conceptions, seeking to show their connection with an old Tractarian idea, called of “doctrine of the simple” by Fogelin, and which we‟ll refer to as the “paradigm of logical elementarity” in this dissertation. We will try to sort out how this paradigm about logical “simplicity” or “elementarity” would have led to a misunderstanding about the kind of “logical generality” involved in several domains, especially the empirical and the phenomenal, as in the intervening period. Our aim is to spot out and examine such misunderstandings with the intention of overcoming them, and thus perhaps offer a proper treatment of language and perception which would allow us to understand the kind of “logical generality” involved in the phenomenal realm, and thus understand what that is meant by “indeterminacy”, “inaccuracy” or “vagueness” in those contexts. / Esta dissertação é uma investigação acerca das noções de “indeterminação”, “inexatidão” e “vagueza”, no contexto da fenomenologia de Wittgenstein exposta no período intermediário do autor. A ocorrência de termos vagos em expressões fenomênicas parece um problema a ser resolvido, desde que ocasione certa insatisfação quanto à falta de “precisão” que deveria ser uma característica de qualquer linguagem adequada. Trataremos desse “problema” acerca da vagueza nas expressões e de algumas reflexões em torno disso, mais especificamente, como elas foram apresentadas no texto Linguagem Fenomenal por Wittgenstein. Exploramos nesta dissertação algumas mudanças no pensamento desse filósofo, desde o Tractatus-Logico-Philosophicus até o texto Linguagem Fenomenal, com a intenção de identificar, esclarecer e, posteriormente, propor, a partir das novas ideias de Wittgenstein ao longo do período intermediário, modos de superar os enganos, denunciados por ele próprio no texto mencionado. Tais enganos, segundo Wittgenstein, seriam os responsáveis por certas concepções ilusórias acerca da “linguagem” e da própria “percepção”. Lidaremos com alguns dos apontamentos feitos por Wittgenstein sobre essas concepções ilusórias, buscando mostrar a conexão dessas com uma antiga ideia tractariana, chamada por Fogelin de “doutrina do simples”, e que nós chamamos nesse trabalho de “paradigma da elementaridade lógica”. Consideraremos, como no período intermediário, esse paradigma acerca da “simplicidade” ou “elementaridade” lógica teria ocasionado uma má compreensão acerca do tipo de “generalidade lógica” envolvida em diversos âmbitos, principalmente os âmbitos linguístico e o fenomênico. Visamos trazer à tona e examinar tais enganos, com a intenção de superá-los, e assim talvez oferecer um tratamento adequado à linguagem e à percepção, que nos permita compreender o tipo de “generalidade lógica” envolvida no âmbito fenomênico, e assim compreender o que se entende por “indeterminação”, “inexatidão” ou “vagueza” nesse contexto.

Η “καθιερωμένη θεώρηση για την πράξη”, όντας δεσμευμένη σε ένα ευρύτερο νατουραλιστικό μοντέλο, κατανοεί την πράξη ως “επιμέρους συμβάν”, το οποίο προκαλείται αιτιακά από συγκεκριμένες νοητικές καταστάσεις. Σ’ αυτό το πλαίσιο, η αιτιακή επίδραση της πρόθεσης υπάγεται στο σύνηθες χιουμιανό μοντέλο της αιτιότητας μεταξύ συμβάντων. Ωστόσο, η εν λόγω θεώρηση αποτυγχάνει ουσιωδώς να ερμηνεύσει τις πράξεις στην εξέλιξή τους, πριν δηλαδή να διαμορφωθεί το απαιτούμενο (από το χιουμιανό μοντέλο) εξατομικευμένο συμβάν. Το γεγονός αυτό έχει ευρύτερες επιπτώσεις για τον τρόπο με τον οποίο αντιλαμβάνεται η εν λόγω προσέγγιση τόσο την έννοια της πρόθεσης, όσο και κατ’ επέκταση το ρόλο του δρώντος. Στην παρούσα εργασία και με αφορμή κυρίως πρόσφατες εργασίες από τους Hornsby και Crowther, επιχειρηματολογώ σχετικά με το ότι μπορούμε να υιοθετήσουμε μια εναλλακτική προσέγγιση, τόσο για την οντολογία της πράξης, όσο και για την πρόθεση, η οποία δίνει ικανοποιητικότερες απαντήσεις στις ανωτέρω προκλήσεις. Κεντρική θέση στα πλαίσια αυτής της προσέγγισης είναι η αναγνώριση της “δραστηριότητας” ως συγκροτησιακού στοιχείου της πράξης, κατά τη διάρκεια της οποίας η αιτιακή συμβολή του δρώντος παραμένει συνεχής, σε συμφωνία με μια αριστοτελικού τύπου προσέγγιση της αιτιότητας. Όι παραδοχές αυτές οδηγούν σε μια θεώρηση της πρόθεσης ως καθοδηγητικής της πράξης καθόλη τη διάρκεια εξέλιξής της. / The “standard story of action” being committed to a broader naturalistic model, understands action as a “particular event”, which is caused by certain mental states. In this context, the causal efficacy of intention is covered by the standard humean model of causality between events. Nevertheless, the story in question substantially fails to give an account for actions as they develop, that is, before the required (by the humean model) individuated event has been formed. That fact has broader effects on the way that the approach in question understands the concept of intention, and ultimately the role of the agent. In this thesis, following mainly on recent papers by Hornsby and Crowther, I argue that we can endorse an alternate approach regarding the ontology of action, as well as intention, that gives more adequate answers to the challenges mentioned above. The main thesis in this context is to recognize “activity” as a constitutional element of action, during which the causal efficacy of the agent remains ongoing, in accordance with an aristotelian type of approach to causality. These commitments result in recognizing intention as guiding action throughout the whole of its development.

[pt] A dissertação relaciona as considerações sobre seguir regras com as críticas ao tratamento extensional do infinito como uma totalidade atual no segundo Wittgenstein. No primeiro capítulo, são apresentadas as críticas de Wittgenstein ao padrão mentalista de solução para determinação do significado, elucidando- se a seguir o que considera-se o cerne do Argumento da Linguagem Privada. A partir disso, argumenta-se que a solução comunitarista, formulada em termos da confirmação de um padrão independente, não pode ser coerentemente atribuída a Wittgenstein. No segundo capítulo, a antinomia histórica entre o infinito pensado como potencial ou como atual é introduzida. São apresentados alguns elementos dos tratamentos propostos por Cantor e Dedekind e, a seguir, as críticas que Wittgenstein faz aos mesmos. No terceiro capítulo, a associação das críticas de Wittgenstein ao tratamento do infinito como uma atualidade extensional com a questão da generalidade lingüística é explorada, e o caráter normativo que Wittgenstein atribui às proposições matemáticas é ressaltado. Mostra-se que, para Wittgenstein, a diferença conceitual entre finito e infinito expressaria a diferença entre contextos empíricos e gramaticais. O infinito, enquanto expressão da generalidade, não poderia ser tratado como passível de descrição extensional sem acarretar confusões, mas só poderia ser pensado no âmbito normativo, enquanto regra. Retorna-se então à questão da determinação de uma regra no contexto da discussão sobre as provas matemáticas e analisa-se a noção de surveyability. No quarto capítulo, considera-se a noção de semelhança de família. Ressalta-se então os pontos centrais da estratégia de Wittgenstein para o problema da determinação de uma regra: o abandono da extensionalidade e da univocidade do significado. Ao invés de classes ou elementos primariamente determinando a relação entre as instâncias de uma regra, teríamos antes a própria relação constituindo tais instâncias, e esta relação interna seria estabelecida na própria prática de emprego linguístico. Finalmente, na conclusão, elabora- se uma reconsideração do que foi desenvolvido a partir das relações entre generalidade e circularidade. / [en] The dissertation relates second Wittgenstein s considerations on following rules with his criticism about the extensional treatment of the infinite as actual totality. In the first chapter, a presentation of Wittgenstein s criticism of the mentalist solution pattern for the determination of the meaning is followed by an account of what is taken to be the heart of the Private Language Argument. On that basis, it is argued that the communitarianist solution, formulated in terms of the confirmation of an independent pattern, cannot be consistently attributed to Wittgenstein. In the second chapter, the historical antinomy between the infinite as potential and the infinite as actual is introduced. Some aspects of Cantor s and Dedekind s approaches, as well as Wittgenstein s criticism of those, are also presented. In the third chapter, the connection between Wittgenstein s criticism of the treatment of the infinite as extensional actuality and the question of linguistic generality is explored, and the normative character attributed by Wittgenstein to mathematical propositions is highlighted. It is shown that, according to Wittgenstein, the conceptual difference between finite and infinite expresses the difference between empirical and grammatical contexts. The infinite as expression of generality could not be treated as susceptible to extensional description without bringing about confusion, and it should, therefore, be thought of solely in a normative context, as a rule. The problem of the determination of a rule is thus retaken in the context of the discussion on mathematical proof, and the notion of surveyability is analysed. In the fourth chapter, the notion of family resemblances is considered. The central aspects of Wittgenstein s approach to the problem of the determination of a rule - the abandonment of extensionality and the univocality of meaning - are emphasised. Instead of classes or elements primarily determining the relationship among the instances of a rule, the relationship itself would constitute such instances, and such internal relationship would be established by linguistic usage itself. Finally, in the conclusion, a general reconsideration of what was discussed is carried out, starting from the relationships between circularity and generality.

The aim was to examine whether measurements of creativity as a domain general trait can explain possible connections between a domain general creative performance and one of the creative domain specific areas, humor. Another aim was to examine whether the maladaptive humor styles can explain possible differences in creative performances. An Internet-based survey was used on 281 students from a Swedish university. The survey was composed of three parts consisting of three separate measuring instruments – Sense of Humor Questionnaire (SHQ-6) was used to measure the respondents’ sense of humor, Humor Style Questionnaire (HSQ) measured the respondents’ humor style and Alternate Uses Test (AUT) gave the respondents an opportunity to account for their creative performances. The results showed no significant correlations between the respondents’ sense of humor and their creative performances. Neither could the maladaptive humor styles explain any significant differences in creative performances. However it was found that a low degree of self-defeating humor style had a significant interaction effect on differences in elaboration between low and high degrees of aggressive humor style.

Emulering och simulering omfattar ett brett spektrum av varierande tekniker på olika nivåer, som involverar både mjuk-, och hårdvaruparadigm. Denna rapport är utformad utifrån ett arkitekturellt perspektiv och ämnar att återge en kortfattad studie av ämnen relaterade till design av en generell referensarkitektur för emulatorer. Särskilt fokus har lagts på sammansättningen av de olika komponenter som vanligtvis utgör en emulator, och dessa ställs sedan i jämförelse med den arkitektur, skapad av Von-Neumann, som idag används i den moderna datorn. Målet är att identifiera de mest lämpade arkitekturella strategierna för att kunna uppnå kvalitetskrav som flexibilitet, generalitet och funktionell lämplighet. Vid skapandet av en prototyp görs en avvägning mellan olika tekniker, vilket resulterar i en hybrid mellan tidskritiska och händelsedrivna paradigm. Vidare jämförs befintliga arkitekturer där modularitet mellan komponenter visar sig vara en återkommande egenskap inom samtliga designer. Det viktigaste resultatet vi kom fram till är att en tillståndsmaskin ofta har en betydande roll inom händelsedrivna system. För den prototyp som designats resulterade detta i en tillståndsmaskin med svagt kopplade deltillstånd som tillsammans leder till aktivering av olika huvudtillstånd, något som går i linje med icke-Von Neumann-alternativ. / The field of emulation and simulation comprises of a broad spectrum of various levels and techniques that spans over both software and hardware paradigms. Written, from mainly an architectural point of view, this paper offers a brief study on sporadic topics related to this broad field which in a wholeness provides relevant and useful insights for the design of a general reference architecture for emulators. Special focus lies on the different components that an emulator comprises of and a parallel is drawn to the physical architecture of the modern computing machine with its well-known Von-Neumann architecture. The goal is to identify best architectural strategies for fulfilling flexibility, generality and functional applicability quality requirements. A narrowed down scope is however kept in mind for a prototype model that will be a hybrid consisting of the event-driven and time-critical paradigms. Different real world architectural examples are compared and modularity with regard to components is identified as a recurrent characteristic in all designs. One of the important findings we made is the inclusion of a state machine component, an important part which allows the event-driven system to function properly. This is with a weak coupled setup of various substates that together activate mainstates, which is in line with alterior thinking to Von Neumann-based systems.

L'objectif de cette thèse a été d'investiguer la question de la généralité et de l'automatisation des processus de contrôle inhibiteur au fil du développement cognitif. Cette question a été approchée par différentes études expérimentales menées chez l'enfant, chez l'adolescent et chez l'adulte. Trois études ont porté sur la généralité des processus de contrôle inhibiteur agissant dans des tâches appartenant à des domaines cognitifs différents. Dans la première, nous avons voulu répondre à deux problématiques liées aux divergences présentes dans la littérature sur le développement du contrôle inhibiteur dans des contextes émotionnels. Le but de cette étude était de déterminer (a) si le contrôle inhibiteur froid (dans des contextes affectivement neutres) et le contrôle inhibiteur chaud (dans des contextes émotionnellement chargés) suivent un pattern développemental identique, particulièrement à l'adolescence et (b) le degré de spécificité de ces deux types de contrôle inhibiteur au fil du développement. Pour cela, nous avons comparé les performances d'enfants de 10 ans, d'adolescents de 13 ans et de jeunes adultes de 21 ans à une tâche de Stroop couleur-mot affectivement neutre et une tâche de Stroop émotionnel. Dans une deuxième étude, nous avons testé auprès d'un groupe d'enfants de 9 ans si les processus inhibiteurs impliqués dans une tâche classique piagétienne de conservation du nombre (i.e., domaine logico-mathématique) étaient identiques ou partiellement identiques à ceux impliqués dans la résolution de la tâche neuropsychologique classique de Stroop couleur-mot (i.e., domaine verbal). Dans une troisième étude, nous avons testé auprès d'un groupe d'enfants de 10 ans et d'un groupe de jeunes adultes de 20 ans la transférabilité des processus inhibiteurs entre une tâche de discrimination de lettres en miroir (dont il a récemment été démontré qu'elle nécessite l'inhibition) et une tâche classique de Stroop couleur-mot, afin de déterminer (a) si la mise en place de mécanismes inhibiteurs impliqués dans la résolution d'un conflit à un niveau perceptif très précoce (la reconnaissance visuelle de symboles) peut faciliter la résolution d'un conflit à un niveau de traitement beaucoup plus tardif (niveau sémantique et moteur), et (b) si l'âge affecte cette transférabilité. Ensuite nous nous sommes intéressés à l'automaticité des processus de contrôle inhibiteur, testant si le contrôle inhibiteur peut fonctionner de manière totalement inconsciente sur des conflits provoqués par deux stimuli subliminaux interférents. Pour cela, nous avons conçu un paradigme d'amorçage composé d'essais dans lesquels un item de Stroop inversé subliminal précédait un item de Stroop visible. Le but de ce paradigme était de vérifier la présence d'un effet d'adaptation de conflit et d'un effet d'amorçage négatif du stimulus subliminal sur le stimulus visible, deux effets rapportés dans les études utilisant des stimuli visibles et suggérant un transfert des processus de contrôle de l'amorce à la cible. Enfin dans une cinquième étude nous avons cherché à savoir si les différences interindividuelles en termes de contrôle inhibiteur découlent en partie des processus prénataux, sur base de l'analyse de la forme sulcale du cortex, considérée comme une caractéristique qualitative de l'anatomie cérébrale déterminée pendant la vie fœtale et stable au cours du développement. En utilisant l'imagerie par résonance magnétique anatomique, nous avons analysé les corrélations entre les performances à une tâche de Stroop couleur-mot et la forme sulcale de deux régions clefs du réseau neuronal du contrôle inhibiteur, le cortex cingulaire antérieur dorsal et le sillon frontal inférieur (qui limite le gyrus frontal inférieur), chez un groupe d'enfants de 10 ans et un groupe d'adultes de 22 ans. En conclusion, cette thèse apporte un nouvel éclairage à la question de la généralité et de l'automatisation des processus d'inhibition exécutive dans une perspective développementale / The objective of this thesis was to investigate the question of the domain-generality of inhibitory control and its automatization throughout development. This question has been approached by various experimental studies in children, adolescents and adults. Three studies have focused on the generality of control processes operating in tasks belonging to different cognitive domains. In the first study, we wanted to answer two questions related to the discrepancies found in the literature on the development of inhibitory control in affectively charged contexts. The aim of this study was to determine (a) whether cool inhibition (control processes in emotionally neutral contexts) and hot inhibition (control processes in emotionally charged contexts) follow the same developmental pattern, and (b) the degree of specificity of these two types of inhibitory control throughout development. We thus compared the performance of 10-year-olds, 13-year-olds and 21-year-olds to an emotionally neutral color-word Stroop task and an emotional Stroop task. In a second study on 9-year-old children, we tested whether the inhibitory processes involved in a classical Piagetian conservation task (i.e., logico-mathematical domain) were identical or partially identical to those involved in the resolution of the classical neuropsychological color-word Stroop task (i.e., verbal domain). In a third study, we tested on a group of 10-year-old children and a group of 20-year-old young adults the transferability of inhibitory processes between a mirror letters discrimination task (that has recently been shown to require inhibition) and a color-word Stroop task, to determine (a) whether the resolution of a conflict at an early perceptual stage of the processing stream (recognition of visual symbols) can facilitate the resolution of a conflict arising at a much later stage (semantic and motor levels), and (b) if age affects this transferability. Then we looked at the automaticity of the inhibition processes by testing whether inhibitory control can operate completely unconsciously on conflicts arising between two subliminal interfering stimuli. To this end, we designed a priming paradigm consisting of trials in which a subliminal reverse Stroop item preceded a visible Stroop item. The aim of this paradigm was to verify the presence of a conflict adaptation effect and a negative priming effect produced by the subliminal items on the visible items. These two effects are commonly reported in studies using visible stimuli and suggest a transfer of control processes from the prime to the probe. Finally, in a fifth study, we investigated whether inter-individual differences in inhibitory control are in part due to prenatal processes, based on the analysis of the sulcal pattern, considered as a qualitative feature of the cerebral anatomy that is determined during fetal life and is stable during development. Using anatomical magnetic resonance imaging, we analyzed the correlations between the performance on a color-word Stroop task and the sulcal pattern of two key regions of the inhibitory control neural network, the dorsal anterior cingulate cortex and the inferior frontal sulcus (which limits the lower frontal gyrus) in a group of 10-year-old children and in a group of 22-year-old adults. In conclusion, this thesis sheds new light on the question of the generality and the automatization of the inhibitory control processes from a developmental perspective.

Managing a system of network services is a complex and large-scale task that often lacks a trivial optimal solution. Deep Reinforcement Learning (RL) has shown great potential in being able to solve these tasks in static settings. However, in practice, the RL agents struggle to generalize their control policies enough to work in more dynamic real-world environments. To achieve a generality between environments, multiple contributions are made by this thesis. Low-level metrics are collected from each node in the system to help explain changes in the end-to-end delay of the system. To achieve generality in its control policy, more ways to observe and understand the dynamic environment and how it changes are provided to the RL agent by introducing the end-to-end delay of each service in the system to its observation space. Another approach to achieving more generality in RL policies is Model-Agnostic Meta-Learning (MAML), a type of Meta-Learning approach where instead of learning to solve a specific task, the model learns to learn how to quickly solve a new task based on prior knowledge. Results show that low-level metrics yield a much greater generality when helping to explain the delay of a system. Applying MAML to the problem is beneficial in adding generality to a learned RL policy and making the adaptation to a new task faster. If the RL agent can observe the changes to the underlying dynamics of the environment between tasks by itself, the policy can achieve this generality by itself without the need for a more complex method. However, if acquiring or observing the necessary data is too expensive or complex, switching to a Meta-Learning approach might be beneficial to increase generality. / Hanteringen av ett system med nätverksstjänster är ett komplext och stor skaligt problem där den optimal lösning inte är trivial. Djup förstärkningsinlärning har visat stor potential i att kunna lösa dessa problem i statiska miljöer. Dock är det svårt att generalisera lösningarna tillräckligt för att fungera i mer komplicerade och realistiska dynamiska miljöer. För att uppnå mer generella lösningar mellan miljöer presenterar denna masteruppsats flera möjliga lösningar. Lågnivåmetrik samlas in från varje nod i systemet för att hjälpa förklara skillnader i den totala responstiden för varje tjänst i systemet. För att generalisera förstärkningsinlärningsmodellen kan den förses med fler sätt att observera miljön, och därmed lära sig förstå hur den dynamiska miljön förändras. En annan metod för att uppnå mer generalitet inom förstärkningsinlärning är Model-Agnostic Meta-Learning (MAML), en typ av Meta-Learning där istället för att lära sig att lösa en specifik uppgift, modellen lär sig att lära sig att snabbt lösa en ny uppgift baserat på sin tidigare kunskap. Resultaten visar att lågnivåmetriken leder till en mycket högre generalitet i att hjälpa till att förklara responstiden av ett system. Att applicera MAML till problemet hjälper att bidra med generalitet till en förstärkningsinlärningsmodell och gör anpassningen till en ny uppgift snabbare. Om modellen själv kan observera ändringarna i underliggande dynamiken bakom miljön mellan uppgifter kan den uppnå mer generalitet utan ett behov av en mer komplex metod som MAML. Däremot, om det är svårt eller dyrt att få tag på eller observera den nödvändiga datan, kan ett byte till en Meta-Learning baserad metod vara mer fördelaktig för att öka generaliteten.

When we develop and use species distribution models to predict species' current or potential distributions, we are faced with the trade-offs between model generality, precision, and realism. It is important to know how to improve and validate model generality while maintaining good model precision and realism. However, it is difficult for ecologists to evaluate species distribution models using field-sampled data alone because the true species response function to environmental or ecological factors is unknown. Species distribution models should be able to approximate the true characteristics and distributions of species if ecologists want to use them as reliable tools. Simulated data provide the advantage of being able to know the true species-environment relationships and control the causal factors of interest to obtain insights into the effects of these factors on model performance. I used a case study on Bythotrephes longimanus distributions from several hundred Ontario lakes and a simulation study to explore the effects on model performance caused by several factors: the choice of predictor variables, the model evaluation methods, the quantity and quality of the data used for developing models, and the strengths and weaknesses of different species distribution models. Linear discriminant analysis, multiple logistic regression, random forests, and artificial neural networks were compared in both studies. Results based on field data sampled from lakes indicated that the predictive performance of the four models was more variable when developed on abiotic (physical and chemical) conditions alone, whereas the generality of these models improved when including biotic (relevant species) information. When using simulated data, although the overall performance of random forests and artificial neural networks was better than linear discriminant analysis and multiple logistic regression, linear discriminant analysis and multiple logistic regression had relatively good and stable model sensitivity at different sample size and detection limit levels, which may be useful for predicting species presences when data are limited. Random forests performed consistently well at different sample size levels, but was more sensitive to high detection limit. The performance of artificial neural networks was affected by both sample size and detection limit, and it was more sensitive to small sample size.

When we develop and use species distribution models to predict species' current or potential distributions, we are faced with the trade-offs between model generality, precision, and realism. It is important to know how to improve and validate model generality while maintaining good model precision and realism. However, it is difficult for ecologists to evaluate species distribution models using field-sampled data alone because the true species response function to environmental or ecological factors is unknown. Species distribution models should be able to approximate the true characteristics and distributions of species if ecologists want to use them as reliable tools. Simulated data provide the advantage of being able to know the true species-environment relationships and control the causal factors of interest to obtain insights into the effects of these factors on model performance. I used a case study on Bythotrephes longimanus distributions from several hundred Ontario lakes and a simulation study to explore the effects on model performance caused by several factors: the choice of predictor variables, the model evaluation methods, the quantity and quality of the data used for developing models, and the strengths and weaknesses of different species distribution models. Linear discriminant analysis, multiple logistic regression, random forests, and artificial neural networks were compared in both studies. Results based on field data sampled from lakes indicated that the predictive performance of the four models was more variable when developed on abiotic (physical and chemical) conditions alone, whereas the generality of these models improved when including biotic (relevant species) information. When using simulated data, although the overall performance of random forests and artificial neural networks was better than linear discriminant analysis and multiple logistic regression, linear discriminant analysis and multiple logistic regression had relatively good and stable model sensitivity at different sample size and detection limit levels, which may be useful for predicting species presences when data are limited. Random forests performed consistently well at different sample size levels, but was more sensitive to high detection limit. The performance of artificial neural networks was affected by both sample size and detection limit, and it was more sensitive to small sample size.