Connectionist Framework Research Articles

Neural Dynamics of Processing Inflectional Morphology: An fMRI Study on Korean Inflected Verbs.

The present study aimed to elucidate the neural mechanisms underpinning the visual recognition of morphologically complex verbs in Korean, a morphologically rich, agglutinative language with inherent polymorphemic characteristics. In an fMRI experiment with a lexical decision paradigm, we investigated whether verb inflection types (base, regular, and irregular) are processed through separate mechanisms or a single system. Furthermore, we explored the semantic influence in processing inflectional morphology by manipulating the semantic ambiguity (homonymous vs. unambiguous) of inflected verbs. The results showed equivalent activation levels in the left inferior frontal gyrus for both regular and irregular verbs, challenging the dichotomy between the two. Graded effects of verb regularity were observed in the occipitotemporal regions, with regular inflections eliciting increased activation in the fusiform and lingual gyri. In the middle occipital gyrus, homonyms showed decreased activation relative to that of unambiguous words, specifically for base and irregular forms. Furthermore, the angular gyrus exhibited significant modulation with all verb types, indicating a semantic influence during morphological processing. These findings support single-system theories and the connectionist framework, challenging the assumptions of purely orthographic morphological decomposition and dual-mechanism accounts. Furthermore, they provide evidence for a semantic influence during morphological processing, with differential reliance on semantic activation for regular and irregular inflections.

The phonological congruency modulated long-term form priming of Chinese characters.

Elucidating the interaction between lexical processing and word learning is essential for a complete understanding of the underlying mechanisms of each of them. Long-term priming for words reflects an interplay between lexical processing and word learning. Although robust long-term priming effects have been found between two occurrences of the same word and between semantically similar words, it remains unclear whether long-term priming between orthographically similar words (i.e., long-term form priming) is a reliable effect. Following the theoretical analysis based on the connectionist framework, we articulated the possibility that long-term form priming might be modulated by the phonological congruency between the prime and target words, and that if this modulator was under control, reliable effects of long-term form priming would emerge. However, this hypothesis has not been adequately tested empirically. The present study tested this hypothesis by using Chinese phonograms and the phonetic radicals embedded in them as the prime and target items. In three experiments that varied in the types of stimuli and testing tasks, we consistently found that when the prime and target had the same phonology, naming the prime facilitated later processing of the target, while when they had different phonologies, the priming effect was inhibitory. These observations were consistent with the connectionist account of long-term priming for words. Our findings help confirm the reliability, generalizability, and robustness of long-term form priming and elucidate its underlying mechanisms, and suggesting promising future directions on the interactions between lexical processing and word learning.

The computational origin of representation.

Each of our theories of mental representation provides some insight into how the mind works. However, these insights often seem incompatible, as the debates between symbolic, dynamical, emergentist, sub-symbolic, and grounded approaches to cognition attest. Mental representations-whatever they are-must share many features with each of our theories of representation, and yet there are few hypotheses about how a synthesis could be possible. Here, I develop a theory of the underpinnings of symbolic cognition that shows how sub-symbolic dynamics may give rise to higher-level cognitive representations of structures, systems of knowledge, and algorithmic processes. This theory implements a version of conceptual role semantics by positing an internal universal representation language in which learners may create mental models to capture dynamics they observe in the world. The theory formalizes one account of how truly novel conceptual content may arise, allowing us to explain how even elementary logical and computational operations may be learned from a more primitive basis. I provide an implementation that learns to represent a variety of structures, including logic, number, kinship trees, regular languages, context-free languages, domains of theories like magnetism, dominance hierarchies, list structures, quantification, and computational primitives like repetition, reversal, and recursion. This account is based on simple discrete dynamical processes that could be implemented in a variety of different physical or biological systems. In particular, I describe how the required dynamics can be directly implemented in a connectionist framework. The resulting theory provides an "assembly language" for cognition, where high-level theories of symbolic computation can be implemented in simple dynamics that themselves could be encoded in biologically plausible systems.

Consciousness as an Emergent Phenomenon: A Tale of Different Levels of Description.

One of the biggest queries in cognitive sciences is the emergence of consciousness from matter. Modern neurobiological theories of consciousness propose that conscious experience is the result of interactions between large-scale neuronal networks in the brain, traditionally described within the realm of classical physics. Here, we propose a generalized connectionist framework in which the emergence of “conscious networks” is not exclusive of large brain areas, but can be identified in subcellular networks exhibiting nontrivial quantum phenomena. The essential feature of such networks is the existence of strong correlations in the system (classical or quantum coherence) and the presence of an optimal point at which the system’s complexity and energy dissipation are maximized, whereas free-energy is minimized. This is expressed either by maximization of the information content in large scale functional networks or by achieving optimal efficiency through the quantum Goldilock effect.

Entrepreneurial dynamism and the built environment in the evolution of university entrepreneurial ecosystems

Abstract University-centered entrepreneurial ecosystems (UCEEs) are complex webs of entrepreneurs, researchers, institutional support structures, and the built environment. We study the relationship between the built environment and the dynamism of the behavior of ecosystem agents in the evolution of UCEEs. Drawing upon data from interviews with ecosystem agents, as well as documents associated with the planning and development of purpose-built facilities (the built environment), we show how planned ecosystem evolution strategies differ from realized strategies. In particular, we develop a model of emergent response mechanisms to ecosystem dynamism and munificence, which include coping, learning, and adapting. We discuss UCEE evolution within a connectionist framework to better address the dynamic interaction of agents, institutions, and the environment.

Associative Memory Model-Based Linear Filtering and Its Application to Tandem Connectionist Blind Source Separation

We propose a blind source separation method that yields high-quality speech with low distortion. Time-frequency (TF) masking can effectively reduce interference, but it produces nonlinear distortion. By contrast, linear filtering using a separation matrix such as independent vector analysis (IVA) can avoid nonlinear distortion, but the separation performance is reduced under reverberant conditions. The tandem connectionist approach combines several separation methods and it has been used frequently to compensate for the disadvantages of these methods. In this study, we propose associative memory model (AMM)-based linear filtering and a tandem connectionist framework, which applies TF masking followed by linear filtering. By using AMM trained with speech spectra to optimize the separation matrix, the proposed linear filtering method considers the properties of speech that are not considered explicitly in IVA, such as the harmonic components of spectra. TF masking is applied in the proposed tandem connectionist framework to reduce unwanted components that hinder the optimization of the separation matrix, and it is approximated by using a linear separation matrix to reduce nonlinear distortion. The results obtained in simultaneous speech separation experiments demonstrate that although the proposed linear filtering method can increase the signal-to-distortion ratio (SDR) and signal-to-interference ratio (SIR) compared with IVA, the proposed tandem connectionist framework can obtain greater increases in SDR and SIR, and it reduces the phoneme error rate more than the proposed linear filtering method.

Effect of Different Teaching Techniques on the Acquisition of Grammatical Gender by Beginning German Second Language Learners

The grammatical gender of German nouns continues to pose a challenge to second language learners. Following from a connectionist framework, this study explores the effect of two input enhancement techniques, color‐coding and gendered actors, on the learning of grammatical gender by beginning learners of German during a vocabulary acquisition activity. Results on an immediate post‐test show no significant differences between conditions. In the delayed post‐test, however, the control group experienced a significant decrease in accurate gender assignment of the target items, while the gendered actor group displayed a non‐significant loss in scores, and the color‐coding group actually displayed slightly better scores compared to the immediate post‐test. These results suggest that color‐coding might be the most effective, while also the least work‐intensive, technique for instructors to implement. Further pedagogical implications resulting from this study are discussed.

Connectionist perspectives on language learning, representation and processing.

The field of formal linguistics was founded on the premise that language is mentally represented as a deterministic symbolic grammar. While this approach has captured many important characteristics of the world's languages, it has also led to a tendency to focus theoretical questions on the correct formalization of grammatical rules while also de-emphasizing the role of learning and statistics in language development and processing. In this review we present a different approach to language research that has emerged from the parallel distributed processing or 'connectionist' enterprise. In the connectionist framework, mental operations are studied by simulating learning and processing within networks of artificial neurons. With that in mind, we discuss recent progress in connectionist models of auditory word recognition, reading, morphology, and syntactic processing. We argue that connectionist models can capture many important characteristics of how language is learned, represented, and processed, as well as providing new insights about the source of these behavioral patterns. Just as importantly, the networks naturally capture irregular (non-rule-like) patterns that are common within languages, something that has been difficult to reconcile with rule-based accounts of language without positing separate mechanisms for rules and exceptions.

Quantum Cybernetics and Complex Quantum Systems Science: A Quantum Connectionist Exploration

Quantum cybernetics and its connections to complex quantum systems science is addressed from the perspective of quantum artificial neural networks as complex quantum computing systems. In this way, the notion of an autonomous quantum computing system is introduced in regards to quantum artificial intelligence, and applied to quantum artificial neural networks, considered as autonomous quantum computing systems, which leads to a quantum connectionist framework within quantum cybernetics for complex quantum computing systems. Several examples of quantum feedforward neural networks are addressed in regards to Boolean functions' computation, multilayer quantum computation dynamics, entanglement and quantum complementarity. The examples provide a framework for a reflection on the role of quantum artificial neural networks as a general framework for addressing complex quantum systems that perform network-based quantum computation, possible consequences are drawn regarding quantum technologies, as well as fundamental research in complex quantum systems science and quantum biology.

Age of acquisition and imageability: a cross‐task comparison

Previous research has reported an imageability effect on visual word recognition. Words that are high in imageability are recognised more rapidly than are those lower in imageability. However, later researchers argued that imageability was confounded with age of acquisition. In the current research, these two factors were manipulated in a factorial design to assess their effect in a lexical decision task and a progressive demasking task. Across both tasks, there was a clear and robust effect of age of acquisition. In contrast, the imageability effect was only evident in the progressive demasking task. Both effects are explained within the connectionist framework in terms of network plasticity and semantic feedback activation.

Cingulum bundle alterations underlie subjective fatigue in multiple sclerosis

Objective: To evaluate the neural basis of subjective fatigue in subjects with multiple sclerosis (MS) using a connectionist framework. Methods: Seventy seven subjects with relapsing–remitting MS were recruited in the study and underwent subjective fatigue evaluations and a diffusion MRI scan. Firstly, local white matter Fractional Anisotropy values were correlated with subjective fatigue scores using a voxel-wise approach. The long-range loss of connectivity due to structural damage in the white matter voxels thus associated with subjective fatigue was then assessed using the Network Modification (NeMo) package. Results: A voxel-wise regression analysis with fatigue scores revealed a significant association between structural damage and fatigue levels in two discrete white matter clusters, both included in the left cingulate bundle. The connectivity analysis revealed that damage in these clusters was associated with loss of structural connectivity in the anterior and medial cingulate cortices, dorsolateral prefrontal areas and in the left caudate. Discussion: Our data point to the cingulum bundle and its projections as the key network involved in subjective fatigue perception in MS. More generally, these results suggest the potential of the connectionist framework to generate coherent models of the neural basis of complex symptomatology in MS.

A latch on priming.

A latch on priming.

Connecting conscious and unconscious processing.

Consciousness remains a mystery-"a phenomenon that people do not know how to think about-yet" (Dennett, 1991, p. 21). Here, I consider how the connectionist perspective on information processing may help us progress toward the goal of understanding the computational principles through which conscious and unconscious processing differ. I begin by delineating the conceptual challenges associated with classical approaches to cognition insofar as understanding unconscious information processing is concerned, and to highlight several contrasting computational principles that are constitutive of the connectionist approach. This leads me to suggest that conscious and unconscious processing are fundamentally connected, that is, rooted in the very same computational principles. I further develop a perspective according to which the brain continuously and unconsciously learns to redescribe its own activity itself based on constant interaction with itself, with the world, and with other minds. The outcome of such interactions is the emergence of internal models that are metacognitive in nature and that function so as to make it possible for an agent to develop a (limited, implicit, practical) understanding of itself. In this light, plasticity and learning are constitutive of what makes us conscious, for it is in virtue of our own experiences with ourselves and with other people that our mental life acquires its subjective character. The connectionist framework continues to be uniquely positioned in the Cognitive Sciences to address the challenge of identifying what one could call the "computational correlates of consciousness" (Mathis & Mozer, 1996) because it makes it possible to focus on the mechanisms through which information processing takes place.

A connectionist model of data compression in memory

Abstract Mental imagery is an integral part of daily life, yet it has been poorly studied. In order to effectively imagine, and make future predictions, it is necessary to obtain an accurate picture of the world. If one were to remember events in complete detail, the memory requirements would be oppressive. Hence, it is necessary to compress the data. In psychology, memory compression has been very poorly studied. On the other hand, in computer technology, video and image compression has been thoroughly researched and standards such as MPEG, JPEG and GIF are largely in use today. We take inspiration from these techniques to form a connectionist framework of data compression. We then apply this framework to a problem in spatial cognition – given the motion of an object in a particular trajectory, its future motion should be predicted. An initial solution to this problem without compression was demonstrated earlier. In this paper, we demonstrate that there is a large memory reduction compared to the earlier system, and that larger simulations that were previously not viable can be run in this system.

Ingenious genes: how gene regulation networks evolve to control development

Each of us is a collection of more than ten trillion cells, busy performing tasks crucial to our continued existence. Gene regulation networks, consisting of a subset of genes called transcription factors, control cellular activity, producing the right gene activities for the many situations that the multiplicity of cells in our bodies face. Genes working together make up a truly ingenious system. In this book, Roger Sansom investigates how gene regulation works and how such a refined but simple system evolved. Sansom describes in detail two frameworks for understanding gene regulation. The first, developed by the theoretical biologist Stuart Kauffman, holds that gene regulation networks are fundamentally systems that repeat patterns of gene expression. Sansom finds Kauffman's framework an inadequate explanation for how cells overcome the difficulty of development. Sansom proposes an alternative: the connectionist framework. Drawing on work from artificial intelligence and philosophy of mind, he argues that the key lies in how multiple transcription factors combine to regulate a single gene, acting in a way that is qualitatively consistent. This allows the expression of genes to be finely tuned to the variable microenvironments of cells. Because of the nature of both development and its evolution, we can gain insight into the developmental process when we identify gene regulation networks as the controllers of development. The ingenuity of genes is explained by how gene regulation networks evolve to control development.

Toward an integrated neurocomputational model of auditory scene analysis

The functional models of auditory scene analysis (ASA) available in the literature have several limitations. First, they independently implement various principles and theories that are specific to the auditory modality. Second, they rarely account for top-down, high level, cognitive effects on ASA. The present paper aims to propose a new integrated model of ASA and reports results within a connectionist modeling framework to account for a wide range of effects on auditory scene analysis. The used connectionist framework is conformed to the known functional and anatomical constraints regarding the biological principles underlying auditory processing. This new neurocomputational model is specifically dedicated to account for top-down effects on ASA, such as attentional control, long-term memory knowledge effects and cross-modal interactions.

Philippe G. Schyns

Philippe G. Schyns

TRACX: A recognition-based connectionist framework for sequence segmentation and chunk extraction.

Individuals of all ages extract structure from the sequences of patterns they encounter in their environment, an ability that is at the very heart of cognition. Exactly what underlies this ability has been the subject of much debate over the years. A novel mechanism, implicit chunk recognition (ICR), is proposed for sequence segmentation and chunk extraction. The mechanism relies on the recognition of previously encountered subsequences (chunks) in the input rather than on the prediction of upcoming items in the input sequence. A connectionist autoassociator model of ICR, truncated recursive autoassociative chunk extractor (TRACX), is presented in which chunks are extracted by means of truncated recursion. The performance and robustness of the model is demonstrated in a series of 9 simulations of empirical data, covering a wide range of phenomena from the infant statistical learning and adult implicit learning literatures, as well as 2 simulations demonstrating the model's ability to generalize to new input and to develop internal representations whose structure reflects that of the items in the input sequence. TRACX outperforms PARSER (Perruchet & Vintner, 1998) and the simple recurrent network (SRN, Cleeremans & McClelland, 1991) in matching human sequence segmentation on existing data. A new study is presented exploring 8-month-olds' use of backward transitional probabilities to segment auditory sequences.

Supplemental Material for TRACX: A Recognition-Based Connectionist Framework for Sequence Segmentation and Chunk Extraction

Supplemental Material for TRACX: A Recognition-Based Connectionist Framework for Sequence Segmentation and Chunk Extraction

Intuition in Judgment and Decision Making: Extensive Thinking Without Effort

We claim that intuition is capable of quickly processing multiple pieces of information without noticeable cognitive effort. We advocate a component view stating that intuitive processes in judgment and decision making are responsible for information integration and output formation (e.g., preference, choice), whereas analytic thinking mainly guides input formation such as search, generation, and change of information. We present empirical evidence corroborating this notion and show that integration of information and preference formation works without cognitive control and is unconstrained by the amount of encoded information and cognitive capacity. We discuss the implications of our findings for the bounded rationality perspective and the multiple strategy approach to judgment and decision making. Finally we outline a connectionist framework for integrating intuitive and analytic thought processes.