Tesi sul tema "Connectionism"

As the technology of computer systems matures, it is becoming clear that conventional techniques are inadequate for complex applications, and attention is being increasingly directed at the use of Knowledge-Based Systems technology. Critical problems in Knowledge-Based Systems are the representation of the expertise from the selected application, and in harnessing sufficient computing power to utilise the stored expertise. An idea which is currently popular in Artificial Intelligence research is that of using parallel processing to ensure that the expertise or knowledge is used effectively, within realistic timescales. Proposals vary in the degree of parallelism, and in the distribution of the problem solving activities. The theory of connectionism, which proposes that the knowledge representation and the problem solving computations are distributed across a very large number of processors, has generated considerable interest and response. Connectionist machines, sometimes known as massively parallel processors, are not highly parallel versions of conventional problem solving engines, but combine the representation of the problem with the processing to produce what is known as an active memory network. Research is described which investigates the application of connectionist theory to various complex problems. These problems are investigated within the context of conventional knowledge-based systems, with the aim of establishing if massively parallel technology realises an efficient problem solving engine. The difficulties associated with the representation and use of numbers in connectionist networks are addressed, the problem of relating the knowledge representation in expert systems to that in connectionist networks is resolved, and the potential for medium scale parallelism in knowledge-based systems is contrasted with the parallelism of connectionism. Finally, the vision of an ideal problem solving engine is considered; some aspects of the evolving designs approach this ideal, and are described.

This thesis presents a connectionist theory of how infinite languages may fit within finite minds. Arguments are presented against the distinction between linguistic competence and observable language performance. It is suggested that certain kinds of finite state automata, i.e recurrent neural networks are likely to have suffcient computational power,and the necessary generalization capability,to serve as models for the procesing and acquisition of linguistic structure.

Connectionism, a new technique for modeling cognitive processes, has been presented by its supporters as a revolutionary advance that will soon replace conventional artificial intelligence (AI) research based on the serial computer. In this thesis, I identify three 'gambits' with which critics attempt to undermine connectionist claims, and show that use of these gambits depends on the status of the respondent's own discipline. I argue that in cases where the respondent's discipline has an accepted identity, for example biology and psychology, they take contradictory stances on the issue of continuity between their discipline and connectionism. By contrast, responses from supporters of AI, which has an uncertain status, insist on a continuous relationship between connectionism and AI. To account for this, I suggest that claims made by both supporters and critics of connectionism, which those actors would regard as purely cognitive, are tacitly structured by Kuhn's model of scientific change. As certain claims which the actors would describe as purely cognitive can be accounted for by the presence in common scholarly parlance of a particular philosophical model of scientific change, I conclude that in the confrontation between connectionism and conventional AI there exists a complex relationship between social and cognitive processes.
Master of Science

The aim of this essay is to explore the potential for an epistemology consistent with eliminative materialism based on work in connectionist modeling.
I present a review of the connectionist approach to psychological models that contrasts it with the classical symbolic approach, focusing on the nature of their respective representations. While defending the legitimacy of the connectionist approach, I find that its most useful application is as a basis for neuroscientific investigation.
Discussing connectionist psychology, I find it inconsistent with folk psychology and therefore consistent with eliminative materialism. I argue also for the naturalization of epistemology and thus for the relevance of psychology for epistemology. The conclusion of the essay is an outline of connectionist epistemology, which centres around two mathematical analyses of the global activity of connectionist networks; I argue that connectionist psychology leads to a version of epistemic pragmatism.

Connectionism's main contribution to cognitive science will prove to be the renewed impetus it has imparted to learning. Learning can be integrated into the existing theoretical foundations of the subject, and the combination, statistical computational theories, provide a framework within which many connectionist mathematical mechanisms naturally fit. Examples from supervised and reinforcement learning demonstrate this. Statistical computational theories already exist for certainn associative matrix memories. This work is extended, allowing real valued synapses and arbitrarily biased inputs. It shows that a covariance learning rule optimises the signal/noise ratio, a measure of the potential quality of the memory, and quantifies the performance penalty incurred by other rules. In particular two that have been suggested as occuring naturally are shown to be asymptotically optimal in the limit of sparse coding. The mathematical model is justified in comparison with other treatments whose results differ. Reinforcement comparison is a way of hastening the learning of reinforcement learning systems in statistical environments. Previous theoretical analysis has not distinguished between different comparison terms, even though empirically, a covariance rule has been shown to be better than just a constant one. The workings of reinforcement comparison are investigated by a second order analysis of the expected statistical performance of learning, and an alternative rule is proposed and empirically justified. The existing proof that temporal difference prediction learning converges in the mean is extended from a special case involving adjacent time steps to the general case involving arbitary ones. The interaction between the statistical mechanism of temporal difference and the linear representation is particularly stark. The performance of the method given a linearly dependent representation is also analysed.

A cognitive process model is developed that predicts the 3 major symbolic comparison response time effects (distance, end, and semantic congruity) found in the results of the linear syllogistic reasoning task. The model assumes that people generate an ordering of a finite set of symbolic stimuli on the basis of information contained in the pairwise relations between adjacent stimulus items. The learning of this ordering is simulated within a simple connectionist framework. The decision-making component of the model utilizes 2 separate evidence accrual processes operating in parallel. One process accumulates information about the positional difference between the stimulus items being compared, and the other accumulates information about the endpoint status of each of those items. A response occurs whenever enough evidence favouring it has been accumulated within either of these processes. The model also assumes that the congruencies between the positions of the stimulus items within the ordering and the form of the comparative instruction can lead to either interfering or facilitating effects on the rate of evidence accumulation within each of these accrual processes. To test the model, data are obtained from the single-session performances of a group of 16 subjects and the multiple-session performances of an additional 2 subjects. The task is a variant of the one used by Trabasso, Riley, and Wilson (1975) and involves paired comparisons of ordered symbolic stimuli (three-letter names). Simulations of the model provide an excellent account of the group mean correct response times, as well as a very good account of the full set of data obtained from the 2 additional subjects (including percentage correct and response time distributional data).

A new computing platform, the General Purpose Reprsentation and Association Machine is studied and simulated. GPRAM machines use vague measurements to do a quick and rough assessment on a task; then use approximated message-passing algorithms to improve assessment; and finally selects ways closer to a solution, eventually solving it. We illustrate concepts and structures using simple examples.
ID: 031001523; System requirements: World Wide Web browser and PDF reader.; Mode of access: World Wide Web.; Adviser: Lei Wei.; Title from PDF title page (viewed August 19, 2013).; Thesis (M.S.E.E.)--University of Central Florida, 2012.; Includes bibliographical references (p. 45).
M.S.E.E.
Masters
Electrical Engineering and Computing
Engineering and Computer Science
Electrical Engineering

Two connectionist models are developed that predict humans' categorization of facial expressions of emotion and their judgements of similarity between two facial expressions. For each stimulus, the models predict the subjects' judgement, the entropy of the response, and the mean response time (RT). Both models involve a connectionist component which predicts the response probabilities and a response generator which predicts the mean RT. The input to the categorization model is a preprocessed picture of a facial expression, while the hidden unit representations generated by the first model for two facial expressions constitute the input of the similarity model. The data collected on 45 subjects in a single-session experiment involving a categorization and a similarity task provided the target outputs to train both models. Two response generators are tested. The first, called the threshold model , is a linear integrator with threshold inspired from Lacouture and Marley's (1991) model. The second, called the channel model, constitutes a new approach which assumes a linear relationship between entropy of the response and mean RT. It is inspired by Lachman's (1973) interpretation of Shannon's (1948) entropy equation. The categorization model explains 50% of the variance of mean RT for the training set. It yields an almost perfect categorization of the pure emotional stimuli of the training set and is about 70% correct on the generalization set. A two-dimensional representation of emotions in the hidden unit space reproduces most of the properties of emotional spaces found by multidimensional scaling in this study as well as in other studies (e.g., Alvarado, 1996). The similarity model explains 53% of the variance of mean similarity judgements; it provides a good account of subjects' mean RT; and it even predicts an interesting bow effect that was found in subjects' data.

This study examines the use of the neural net paradigm as a modeling tool to represent word meanings. The neural net paradigm, also called "connectionism" and "parallel distributed processing," provides a new metaphor and vocabulary for representing the structure of the mental lexicon. As a research method applied to the componential analysis of word meanings, the neural net approach has one primary advantage over the traditional introspective method: freedom from the investigator's personal biases.The connectionist method is illustrated in this thesis with an extensive examination of the meanings of the words "cup" and "mug." These words have been studied previously by Labov (1973), Wierzbicka (1985), Andersen (1975), and Kempton (1978), using very different methods.The neural net models developed in this study are based on empirical data acquired through interviews with nine informants who classified 37 objects, 37 photographs, and 37 line drawings as "cups," "mugs," or "neither." These responses were combined with a data file representing the coded attributes of each object, to construct neural net models which reflect each informant's classification process.In the neural net models, the "cup" and "mug" features are interconnected with positive and negative weights that represent the association strengths of the features. When the connection weights are set so that they reflect the informants' responses, the neural net models can account for the extreme discrepancies in object-naming among informants, and the models can also account for the inconsistent classifications of each individual informant with respect to the mode of presentation (drawing, photograph, or actual object). Further, the neural net modelscan predict classifications for novel objects with an accuracy varying from 82% to 100%.By examining the connection weight patterns within the neural net model, it is possible to discover the "cup" and "mug" features which are most salient for each informant, and for the informants collectively. This analysis shows that each informant has acquired internal meanings for the words "cup" and "mug" which are unique to the individual, although there is considerable overlap with respect to the most salient features.
Department of English

Orientador: Edson Françozo
Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem
Made available in DSpace on 2018-08-07T04:10:15Z (GMT). No. of bitstreams: 1 Garcia_RicardoBasso_M.pdf: 578837 bytes, checksum: c5ae075736e4e6e2077a3d8f82d9be89 (MD5) Previous issue date: 2006
Resumo: Esta dissertação começa com uma reflexão sobre o caráter interdisciplinar da psicolingüística, que é tomada como o campo de pesquisa que investiga, entre outros, os processos cognitivos subjacentes ao comportamento lingüístico. Essa investigação emprega diferentes métodos e técnicas, dada a complexidade do estudo de processos que são internos ao organismo e, portanto, não diretamente observáveis. Nesse quadro, introduzimos a metodologia adotada nesta dissertação - a modelagem. O uso de modelos é tratado em termos gerais, para então ser apresentado no contexto do estudo dos processos cognitivos. Neste ponto, é crucial notar que modelar esses processos implica em fazer suposições sobre como eles são. A hipótese geral que fundamenta os modelos de nosso interesse é a de que cognição é computação, ou seja, processos cognitivos são computacionais. Essa hipótese deu origem a dois paradigmas de modelagem distintos - um baseado nos fundamentos da computação digital e outro nos princípios da neurocomputação. Este segundo é apresentado em detalhes, depois de uma exposição das principais diferenças entre esses paradigmas. Estabelecidas as bases teórico-metodológicas dessa abordagem, replicamos o experimento de Elman (1990) em que uma rede neural recorrente simples é treinada para processar palavras dispostas em seqüências que compõem sentenças simples. É interessante notar que essa disposição reflete informações gramaticais, que podem ser descritas em tennos sintáticos (como sujeito, verbo e objeto) e semânticos (como agente, paciente e tema). Nosso intuito foi investigar se e como informações desse tipo são aprendidas pelo sistema. A análise dessa rede mostrou que seu processamento opera principalmente por distinções lexicais, isto é, não há o uso efetivo de informações lingüísticas do nível de sentenças. O passo seguinte foi modificar esse experimento - treinamos a rede com o mesmo conjunto de sentenças, desta vez marcando a fronteira entre elas. A análise dessa rede mostrou que, nesse novo experimento, o processamento opera principalmente por distinções sintáticas, havendo também a presença de distinções semânticas de papel temático. A partir desses resultados, elaboramos um novo experimento - treinamos uma rede com um conjunto de sentenças nas quais a disposição das palavras reflete principalmente informações semânticas de papel temático, havendo também a presença de informações sintáticas e da fronteira entre as sentenças. O objetivo foi controlar a informação lingüística presente nas sentenças, de modo a permitir maior confiabilidade nos resultados. A análise dessa rede mostrou-se coerente com a do experimento anterior, ou seja, o processamento está, de fato, operando segundo conhecimentos sintático-semânticos do nível de sentenças. A consistência dos resultados obtidos permite afirmar, com um grau de confiabilidade satisfatório, que há a presença de conhecimento sintático-semântico, durante o processamento de sentenças, ao mesmo tempo em que mostra como essas informações estão presentes. Esses resultados são importantes porque estendem a já conhecida capacidade da rede recorrente simples em codificar informação sintática, mostrando também sua capacidade em captar informação semântica
Abstract: This dissertation begins with a retlection about the interdisciplinary roots of psycholinguistics, which is taken as the research field concemed with the investigation of the cognitive processes subjacent to the verbal behavior. This investigation employs different methods and techniques, given the complexity inherent to the study of processes that take place inside the organism, and therefore are not directly observable. In view of this, we introduce the methodology adopted in this dissertation - modeling. The usage of models is initially dealt with in general terms, and then is analysed in the context of the study of cognitive processes. At this point, it is crucial to notice that modeling such processes implies making assumptions about how they are. The general hypothesis underlying the models we are interested in is that cognition is computation, that is, cognitive processes are computational. This hypothesis has given birth to two different paradigms of modeling - one based on the fundamentals of digital computing, and other based on the principIes of neurocomputing. Afier a brief discussion of their main differences, the later is presented in details. Once the theoretical-methodological basis of the neurocomputing framework is established, we replicate Elman's (1990) experiment in which a simple recurrent neural network is trained to process words in sequences, forming simple sentences. It should be noted that the sequences reflect grammatical information that can be described in syntactic (subjects, verbs and objects) and semantic (agents, patients and themes) terms. Our intent was to investigate whether and how such information is leamed by the system. Analyses of this experiment have shown that the processing is guided by lexical information, that is, linguistic information at the sentence leveI is not used. In the next step we have modified this experiment by training the network with the saroe set of sentences, just introducing an end-of-period mark between them. The analyses have shown that, in this new experiment, the processing is guided by both syntactic and semantic information, such as thematic roles. Taking these results as starting point, we devised a new experiment - a similar network was trained with a set of sentences in which word sequences were main1y led by semantic information like thematic role; the end-of-period mark between sentences was used. The intent was to controllinguistic information so as to increase our confidence in the results. Results consistent with the fonner experiments were obtained, that is, the processing is indeed guided by both syntactic and semantic infonnation at the sentence level. The consistency of the results achieved a1lows us to state that there are both syntactic and semantic knowledge operating during sentence processing. This result is important because it extends the already known simple recurrent network ability in coding syntactic information, showing that SRNs also are capable of handling semantic infonnation like thematic roles
Mestrado
Mestre em Linguística

The purpose of the present study was to establish evidence for the suggested integration of the theories of connectionism and leadership. Recent theoretical writings in the field of leadership have suggested that the dynamic representations generated by the connectionist perspective is an appropriate approach to understanding how we perceive leaders. Similarly, implicit leadership theory (ILT) explains that our cognitive understandings of leaders are based on a cognitive structure that we use as a means of understanding and categorizing the behaviors of individuals we believe to be leaders. It was predicted that when asked to select a leader from a group of potential leaders, individuals select the leader based on personal belief alignment when the context of the leader selection is personally relevant, or based on cognitive expectations when the context is low in personal relevance. In addition, when experiencing moments of greater personal relevance, individuals will experience a more dynamic cognitive representation of a leader that those experiencing the moment as less personally relevant. Sixty-seven individuals provided usable data from a repeated measures design that asked participants to record their cognitive representations of a leader, participate in a leader selection task, and provide information about their cognitive representations of a leader after the selection task. The results of the study provide support the expectations of the experimenter and the suggestions of the connectionist perspective.
Master of Science

In this thesis we have investigated what concepts are and how they may be represented. We have seen that conceptual representations can be achieved by employing distributed representations in a hidden layer of a neural net- work. A pattern of activity is in this respect a conceptualization while the concept(s) it belongs to is a region of space treated alike by similarity based generalization. That is, the conceptualization may still have its individual properties only attributed to itself, but the properties relevant to the concept are shared among the representations in that region of space. These regions of space are allocated as dictated by coherently covarying properties of the domain, and thus constitutes a hierarchical representation of it. In this hier- archical representation, the most general concepts occupy the largest amount of space, with their subordinate concepts distributed in clusters allocated inside this space. This hierarchic representation is discovered in a coarse to fine manner, mirroring the conceptual development of a child. Properties being highly typical for a concept are, however, easier to learn and may thus be acquired before properties of concepts superordinate to them, mirroring basic level advantages in lexical acquisition. These typical properties show a higher level of activation throughout training. Frequency of presentation also influences how easy a concept or pattern is to acquire. Frequency of presentation causes a higher pressure to differentiate the instance, thus allo- cating a larger amount of space to it. This in turn facilitates the learning of its individual properties, thus attenuating the basic level advantages. The properties that covary coherently in the domain becomes more salient than other properties. This allows concepts to be acquired based on especially in- formative properties, thus possibly overlooking perceptual similarity. When noise was introduced into the system, the hierarchy broke down in a fine to coarse manner. These effects are all due to similarity based generalization and the coarse to fine differentiation of conceptual distinctions, and support many findings in semantic cognition. PDP thus serve as a good starting place for achieving conceptual representations. By viewing concepts as simulators (Barsalou, 1999; Barsalou, 2003a; Barsalou, 2003b), they are a skill to produce context-specific representations. This is also true of the hidden layer conceptual representations, although depending on whether the context is predictive. A simulator is comprised by a set of modality specific perceptual symbols extracted from perceptual states. Barsalou (1999) also offered valuable insights as to how simulators can support productivity and abstract thought. We have also seen how categorization can influence perceptual discrim- ination (Goldstone, 1994). By acquiring categories, the category relevant boundaries acquire distinctiveness with emphasis on the category boundary. For separable dimensions, the irrelevant dimension may receive acquired acquired similarity, however, one null effect was also found in (Goldstone, 60 1994). For Integral dimensions, the irrelevant dimension also acquired dis- tinctiveness. When two dimensions were relevant for categorization, the separable dimensions competed with each other, while the integral did not. Based on results from (Gluck & Meyers, 1993) I have proposed that a predic- tive auto-encoder can account for the results found for separable dimensions. During categorization learning, the stimuli along with the assigned category is processed by a predictive auto-encoder. The result is that predictive dimensions acquire distinctiveness while redundant ones acquire similarity. Whether or not the irrelevant dimension acquire similarity will thus depend on whether it has previously been predictive. Language is another factor influencing perceptual discrimination. When language was introduced into the system it had a profound influence on the conceptual representations (Cangelosi & Parisi, 2001). The representations acquired within category similarity and between category distinctiveness. The effect was largest for verbs, but was also present during non-linguistic processing. Language thus helped the network perfect its conceptual skills with respect to non-linguistic behavior. In Cangelosi & Riga (2006) language was used to implement grounding transfer. This is a process where new behavior is acquired by grounding it in previously learned behavior. This was achieved in the guid- ance of language. This could also be seen as an implementation of Barsalou’s (1999) productivity mechanism. The involvement of language in simulating abstract thought has also been discussed. With reference to cangelosi & Parisi (2001) and Cangelosi & Riga (2006) it seems that language has a profound effect on conceptual processing. Dimensionality of the representation is another important factor in con- ceptual representations. As the dimensionality increases, the number of examples necessary to reach a given level of performance increases exponen- tially (Edelman & Intrator, 1997). Auto-encoders is a common method for unsupervised dimensionality reductions which also preserves the topology of the original domain. The dimensionality is reduced by compressing redun- dant information thus allowing conception to focus on the relevant aspect of the representation. We have also reviewed a theory if prefrontal cortex function suggesting its implication in guiding computation along processing specific pathways and also in acquiring categories and rules (Miller & Freedman, et. al., 2002; Miller & Cohen, 2001; Braver & Cohen, 2000). The PFC thus seems es- sential in conception. However, as the rules learned in the PFC is executed frequently, they get “pushed” down to more autonomous areas of the brain and thus become more autonomous. The PFC will thus be most involved in behavior requiring attention, among which acquiring concepts certainly belongs. A framework for higher level cognitive behavior from Veflingstad & Yildirim (2007) was introduced. This framework was introduced within three levels of cognition: the stimulus-resonse level, the conceptual level 61 and the language level. Within this framework it is proposed that algo- rithms exist in the brain and that they are represented non-symbolically at the conceptual level. They operate on non-symbolic concepts and makes decisions using feed forward networks modeling an if-then rule. By em- ploying distributed representations these algorithms exhibit the properties we have this far discussed and will thus exhibit semantic task performance. These algorithms help experiencing more complex thought and are engaged in higher level cognitive tasks such as planning. A simulation of a non- symbolic summation algorithm was presented showing the feasibility of the approach. It was proposed that the PFC is in charge of learning these algo- rithms, but as they are frequently executed they get “pushed” down to more autonomous areas of the brain and thus no longer require as much attention to be executed. Novelty was proposed as a means of autonomous exploration and a con- tinuous “type checking” parameter. Novelty is an informative and important “signal” as it allows one to assess knowledge of a perceived instance without any explicit reference of memory. This was implemented in a simulation as the sum of differences between the input pattern and the output pattern of an auto-encoder. The simulation showed that novelty could be reliably as- sessed within and between modalities as long as the environment was noise free. When noise was introduced, the performance dropped. The simula- tion was, however, very constrained as the link between the modalities only supported “one to one” relationships. It was therefor suggested that novelty of associations was better assessed as the amount of selective attention the PFC must exert in order for a pattern of activity in a massively recurrent system to settle into a new attractor. It should be mentioned that novelty is here interpreted very broadly. It might be possible that a specific association has been observed many times but that some other association overrides it in the system. This association would thus not be novel in that it has not been experienced but in that it has not been learned to a sufficient degree. Novelty is here also used as an assessment of which of two associations are least familiar. Novelty in this respect would thus be a measure of the amount of stress a current line of processing introduces in the system. From the material presented in this thesis I will in line with Barsalou (2003b) conclude that the concept arises from a skill for producing context- specific representations. This skill arises from interacting with the world and observing meaningful relationships and properties within it. As this skill improves, perception is affected in a way further facilitating this skill. Once this skill has reached a certain level, language can be acquired, improving this skill even more. This in turn, probably facilitates further acquisition of language. Within reference to the three levels proposed there seems to be a circular dependency between the layers with the concept arising from this interaction. However, since conception can arise simply by similarity based generalization, language would not seem necessary for conception. It 62 does seem important in the complex conceptual abilities to humans though. Even though it is here concluded that the concept emerges from the skill of the system, this does not mean that it can not be investigated as patterns of activation. As wee have seen, much can be learned from these patterns. They can also be employed in algorithms achieving more complex thought.

In this work we explore how close the artificial intelligence community has come to model the human mind regarding representation and processing of sequences. We analyse results produced by cognitive psychologists, who explore real minds, for features exhibited by human short- and long-term memory when representing and processing sequences. We compare these features with theories and models from the AI community divided into two types of theories: intrinsic and extrinsic theories. We conclude that the intrinsic theories have managed to explain most of the features, whereas the extrinsic theories still have a lot to do before exhibiting all features. We also present several suggestions for continued research to the AI community within the area of sequence representation and processing in the human mind.

Orientador: Marcos Antonio Alves
Banca: Osvaldo Frota Pessoa Júnior
Banca: Max Rogério Vicentini
Resumo: Nesta Dissertação desenvolvemos uma análise do estatuto científico da ciência cognitiva, em sua fase inicial, mais especificamente no período entre as décadas de 1940 e 1970. Como ponto de partida, utilizamos a abordagem epistemológica de Thomas Kuhn sobre as revoluções científicas. Para alcançar esse objetivo, dividimos a dissertação em três capítulos. No primeiro, expomos alguns dos principais conceitos relacionados à abordagem de Kuhn referentes à estrutura das revoluções científicas. Um dos conceitos-base dessa perspectiva é a noção de paradigma. Um paradigma estabelece e norteia a atividade de uma comunidade científica. Ele é constituído, dentre outras coisas, por teorias empiricamente testáveis, métodos de pesquisa, experimentos, formas de procedimentos, conjuntos de leis e princípios. Indica, ainda, os problemas a serem desenvolvidos pela comunidade científica, determinando uma agenda científica. Via de regra, a fase inicial de uma nova área de pesquisa é marcada por um momento de luta paradigmática, caracterizando-se pela existência de diversos paradigmas rivais. Desde o momento em que um deles passa a ser dominante, a área de pesquisa alcança o estatuto de ciência normal. Uma vez apresentada, de maneira geral, a abordagem de Kuhn, no segundo capítulo passamos a tratar da ciência cognitiva em sua fase inicial. Ainda com raízes na cibernética, esse movimento intelectual, em seus primeiros momentos, almejava instaurar uma ciência dos processos cognitivos. Essa perspecti... (Resumo completo, clicar acesso eletrônico abaixo)
Abstract: In this dissertation, we seek to develop an analysis of the scientific status of cognitive science in its initial phase, more specifically the period between the 1940s and 1970s. As a starting point for this analysis, we use Thomas Kuhn's epistemological approach to revolutions scientific research. To achieve this goal, we divided the dissertation into three chapters. In the first, we present some of the main concepts related to Kuhn's approach to the structure of scientific revolutions. One of the basic concepts of this perspective is the notion of paradigm. A paradigm establishes and guides the activity of a scientific community. It is constituted, among other things, by empirically testable theories, methods of research, experiments, forms of procedures, sets of laws and principles. It also indicates the problems to be developed by the scientific community, determining a scientific agenda. As a rule, the initial phase of a new area of research is marked by a moment of paradigmatic struggle, characterized by the existence of several rival paradigms. From the moment one of them becomes dominant, the area of research reaches the status of normal science. Having presented, in general, Kuhn's approach, in the second chapter we turn to cognitive science in its initial phase. Still rooted in cybernetics, this intellectual movement, in its first moments, aimed to establish a science of cognitive processes. This perspective adopts by methodological principle the conception that cer... (Complete abstract click electronic access below)
Mestre

This thesis puts forward a formal theory of levels and algorithms to provide a foundation for those terms as they are used in much of cognitive science and computer science. Abstraction with respect to concreteness is distinguished from abstraction with respect to detail, resulting in three levels of concreteness and a large number of algorithmic levels, which are levels of detail and the primary focus of the theory. An algorithm or ideal machine is a set of sequences of states defining a particular level of detail. Rather than one fundamental ideal machine to describe the behaviour of a complex system, there are many possible ideal machines, extending Turing's approach to reflect the multiplicity of system descriptions required to express more than weak input-output equivalence of systems. Cognitive science is concerned with stronger equivalence; e.g., do two models go through the same states at some level of description? The state-based definition of algorithms serves as a basis for such strong equivalence and facilitates formal renditions of abstraction and implementation as relations between algorithms. It is possible to prove within the new framework whether or not one given algorithm is a valid implementation of another, or whether two unequal algorithms have a common abstraction, for example. Some implications of the theory are discussed, notably a characterisation of connectionist versus classical models.

Made available in DSpace on 2014-06-12T23:01:00Z (GMT). No. of bitstreams: 2 arquivo8618_1.pdf: 1604073 bytes, checksum: d9765092ab2a50cd8b1426f73b5ed129 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2006
Este estudo é uma revisão narrativa da literatura sobre os paradigmas computacionais, as modelagens naturais conexionistas e a investigação dos fenômenos psicopatológicos. O objetivo geral foi realizar uma coleta de informações sobre os trabalhos publicados, até então, que contemplassem os modelos de processamento de informações no cérebro humano, a analogia com Redes Neurais Artificiais e a aplicação de métodos investigativos nas psicopatologias. A seleção dos estudos foi baseada principalmente pesquisas em bancos de dados digitais: Medline, Períodos CAPES, MIT Search, Scholar Google e PsychInfo, usando os descritores neural networks, neurocomputation, psychopathology, connectionism, mood disorders, depression, cognition e artificial intelligence, em mecanismos de busca digital. Foram selecionados os estudos considerando os critérios de inclusão a partir dos descritores, o aspecto cronológico, a adequação e pertinência dos estudos e o impacto destes artigos na comunidade científica. A literatura clássica também foi incluída. O estudo buscou estabelecer relações entre as pesquisas que utilizaram ferramentas computacionais, visando a criação de modelos que simularam o funcionamento cognitivo do cérebro humano. Destes modelos, as Redes Neurais Artificiais Conexionistas (RNA) mostraram-se as mais promissoras dentre as demais. Conclui-se que as investigações dos fenômenos psicopatológicos baseadas em modelagem computacional conexionista constituem em uma importante estratégia para compreensão do funcionamento da mente humana e de como se processam as alterações psíquicas

This work discusses a way of thinking the normative practices as a phenomenon better understood through a pragmatic account of social practices. We claim that an appropriate approach to normativity should accept the presence, in the normative creature, of natural capacities intrinsically related to norm-governed activities, especially what we call a normative attitude. Thus, we present a discussion on the rule-based account of normativity understood as a sort of intersubjective practice grounded in practical skills and learning processes as well. We also indicate why the pragmatic model appropriately fits with a connectionist model of cognition. Finally, we argue that normative practices should be understood primarily in terms of internal patterns, functionally defined, instituted as nonexplicit, non-conscious individual processes. The consequence is a practical, inferentialist, connectionist, and implicit approach to the normativity.
Este trabalho discute um modo de pensar as práticas normativas como um fenômeno melhor entendido através de uma explicação pragmática das praticas sociais. Afirmamos que uma estratégia apropriada para entender a normatividade deve aceitar a presença, nas criaturas normativas, de capacidades naturais intrinsecamente relacionadas às atividades governadas por normas, especialmente o que chamamos de atitudes normativas. Assim, apresentamos uma discussão de uma abordagem da normatividade baseada em regras entendidas como um tipo de prática intersubjetiva fundada em habilidades práticas e, também, em processes de aprendizado. Indicamos, então, por que um modelo pragmático se adéqua apropriadamente a um modelo conexionista de cognição. Finalmente, argumentamos que as práticas normativas devem ser entendidas primeiramente em termos de padrões internos, funcionalmente definidos, instituídos como processos individuais não explícitos e não conscientes. A consequência é uma explicação prática, inferencialista, conexionista e implícita da normatividade.

Bodily Sensation in Contemporary Extreme Horror Film provides a theory of horror film spectatorship rooted in the physiology of the viewer. In a novel contribution to the field of film studies research, it seeks to integrate contemporary scientific theories of mind with psychological paradigms of film interpretation. Proceeding from a connectionist model of brain function that proposes psychological processes are underpinned by neurology, this thesis contends that whilst conscious engagement with film often appears to be driven by psychosocial conditions – including cultural influence, gender dynamics and social situation – it is physiology and bodily sensation that provide the infrastructure upon which this superstructure rests. Drawing upon the philosophical works of George Lakoff, Mark Johnson and Alain Berthoz, the argument concentrates upon explicating the specific bodily sensations and experiences that contribute to the creation of implicit structures of understanding, or embodied schemata, that we apply to the world round us. Integrating philosophy with contemporary neurological research in the spheres of cognition and neurocinematics, a number of correspondences are drawn between physiological states and the concomitant psychological states often perceived to arise simultaneously alongside them. The thesis offers detailed analysis of a selection of extreme horror films that, it is contended, conscientiously incorporate the body of the viewer in the process of spectatorship through manipulation of visual, auditory, vestibular, gustatory and nociceptive sensory stimulations, simulations and the embodied schemata that arise from everyday physiological experience. The phenomenological film criticism of Vivian Sobchack and Laura U. Marks is adopted and expanded upon in order to suggest that the organicity of the human body guides and structures the psychosocial engagement with, and interpretation of, contemporary extreme horror film. This project thus exposes the body as the architectural foundation upon which conscious interaction with film texts occurs.

The paper discusses the basics of Cognitive Science and describes the achievements of research at the field of Artificial Intelligence
Die im Rahmen des Seminars "Language - Mind - Brain: An Introduction to Psycholinguistics" der englischen Sprachwissenschaft entstandene Arbeit befasst sich mit den Grundlagen der Kognition und diskutiert die Ergebnisse der Forschung auf dem Gebiet der Künstlichen Intelligenz