Belief networks are directed acyclic graphs in which the nodes represent propositions (or variables), the arcs signify direct dependencies between the linked propositions, and the strengths of these dependencies are quantified by conditional probabilities. A network of this sort can be used to represent the generic knowledge of a domain expert, and it turns into a computational architecture if the links are used not merely for storing factual knowledge but also for directing and activating the data flow in the computations which manipulate this knowledge. The first part of the paper deals with the task of fusing and propagating the impacts of new information through the networks in such a way that, when equilibrium is reached, each proposition will be assigned a measure of belief consistent with the axioms of probability theory. It is shown that if the network is singly connected (e.g. tree-structured), then probabilities can be updated by local propagation in an isomorphic network of parallel and autonomous processors and that the impact of new information can be imparted to all propositions in time proportional to the longest path in the network. The second part of the paper deals with the problem of finding a tree-structured representation for a collection of probabilistically coupled propositions using auxiliary (dummy) variables, colloquially called "hidden causes. " It is shown that if such a tree-structured representation exists, then it is possible to uniquely uncover the topology of the tree by observing pairwise dependencies among the available propositions (i.e., the leaves of the tree). The entire tree structure, including the strengths of all internal relationships, can be reconstructed in time proportional to n log n, where n is the number of leaves.

We develop a connectionist approach to processing in quasi-regular domains, as exemplified by English word reading. A consideration of the shortcomings of a previous implementation (Seidenberg & McClelland, 1989, Psych. Rev.) in reading nonwords leads to the development of orthographic and phonological representations that capture better the relevant structure among the written and spoken forms of words. In a number of simulation experiments, networks using the new representations learn to read both regular and exception words, including low-frequency exception words, and yet are still able to read pronounceable nonwords as well as skilled readers. A mathematical analysis of the effects of word frequency and spelling-sound consistency in a related but simpler system serves to clarify the close relationship of these factors in influencing naming latencies. These insights are verified in subsequent simulations, including an attractor network that reproduces the naming latency data directly in its time to settle on a response. Further analyses of the network's ability to reproduce data on impaired reading in surface dyslexia support a view of the reading system that incorporates a graded division-of-labor between semantic and phonological processes. Such a view is consistent with the more general Seidenberg and McClelland framework and has some similarities with---but also important differences from---the standard dual-route account.

. Low-dimensional representations of sensory signals are key to solving many of the computational problems encountered in high-level vision. Principal Component Analysis has been used in the past to derive practically useful compact representations for different classes of objects. One major objection to the applicability of PCA is that it invariably leads to global, nontopographic representations that are not amenable to further processing and are not biologically plausible. In this paper we present a new mathematical construction---Local Feature Analysis (LFA)---for deriving local topographic representations for any class of objects. The LFA representations are sparse-distributed and, hence, are effectively low-dimensional and retain all the advantages of the compact representations of the PCA. But unlike the global eigenmodes, they give a description of objects in terms of statistically derived local features and their positions. We illustrate the theory by using it to extract loca...

A micro-world is described, in which many analogies involving strikingly different concepts and levels of subtlety can be made. The question "What differentiates the good ones from the bad ones?" is discussed, and then the problem of how to implement a computational model of the human ability to come up with such analogies (and to have a sense for their quality) is consicered. A key part of the proposed system, now under development, is its dependence on statistically emergent properties of stochastically interacting "codelets" (small pieces of ready-to-run code created by the system, and selected at random to run with probability proportional to heuristically assigned "urgencies"). Another key element is a network of linked concepts of varying levels of "semanticity", in which activation spreads and indirectly controls the urgencies of new codelets. There is pressure in the syste.rn toward maximizing the degree of "semanticity" or "intensionality" of descriptions of structures, but many such pressures, often conflicting, m.,u$ interact with one another, and compromises must be made. The shifting of (1) percei/v'bd boundaries insi,d,e stru,c, tures, (2) descriptive concepts chosen to apply to structures, and (3)4eatures perceived as salient or not, is called "slippage". What can slip, and how, are emergent consequences of the interaction of (1) the temporary ("cytoplasmic") structures involved in the analogy with (2) the permanent ("Platonic") concepts and links in the conceptual proximity network, or "slippability network". The architecture of this system is postulated as a general architecture suitable for dealing not only with fluid analogies, but also with other types of abstract perception and categorization tasks, such as musical perception, scientific theorizing,...

What is the neural representation of a speech code as it evolves in time? A neural model simulates data concerning segregation and integration of phonetic percepts. Hearing two phonetically related stops in a VC-CV pair (V = vowel; C = consonant) requires 150 ms more closure time than hearing two phonetically different stops in a VC,-C2V pair. Closure time also varies with long-term stimulus rate. The model simulates rate-dependent category boundaries that emerge from feedback: interactions between a working memory for short-term storage of phonetic items and a list categorization network for grouping sequences of items. The conscious speech code is a resonant wave. It emerges after bottom-up signals from the working memory select list chunks which read out top-down expectations that amplify and focus attention on consistent working memory items. In VCi-C2V pairs, resonance is reset by mismatch of Cj with the C, expectation. In VC-CV pairs, resonance prolongs a repeated C. What is the nature of the process that converts brain events into behavioral percepts? An answer to this question is needed in order to understand how the brain controls behavior and how the brain is, in turn, shaped by environmental feedback that is experienced on the behavioral level. The nature of this connection also needs to be understood in order to develop neurally plausible connectionist models. Without it, a correct linking hypothesis cannot be developed between psychological data and the brain mechanisms from which they are generated.

this paper, we describe GENET, a generic neural network simulator, that can solve general CSPs with finite domains. GENET generates a sparsely connected network for a given CSP with constraints C specified as binary matrices, and simulates the network convergence procedure. In case the network falls into local minima, a heuristic learning rule will be applied to escape from them. The network model lends itself to massively parallel processing. The experimental results of applying GENET to randomly generated, including very tight constrained, CSPs and the real life problem of car sequencing will be reported and an analysis of the effectiveness of GENET will be given. NETWORK MODEL The network model is based on the Interactive Activation model (IA) with modifications to suit the natures of the CSPs as defined at the beginning of this paper. The IA model in its original form can be characterized as weak constraint satisfaction, in which the connections represent the coherence, or compatibility, between the connected nodes. This model was developed for associative information retrieval or pattern matching [11, 12]. However, it is not adequate for solving CSPs in general, for which all the constraints are absolute and none of them should be violated at all. For this purpose, the following modifications have been developed. 1. The nodes in the network are grouped into clusters with each cluster representing a variable in Z, and the nodes in each cluster represent the values that can be assigned to the variable. 2. Only inhibitory connections are allowed. The inhibitory connections represent the constraints that do not allow the connected nodes to be active (i.e. turned on) simultaneously. 3. The nodes in the same cluster compete with each other in convergence cycles. The node...

nce is often relative rather than absolute. We conclude that learning to read results in the progressive development of an inferotemporal region increasingly responsive to visual words, which is aptly named the visual word form area (VWFA). D 2004 Elsevier Inc. All rights reserved. Keywords: Word recognition; Occipitotemporal; Specialization Introduction The efficiency of reading in literate adults rests on the ability to quickly identify visual words across large variations of irrelevant parameters such as position, size, color, font, or case. This perceptual expertise requires no less than 5 years of academic training in a specific writing system (Aghababian and Nazir, 2000). The outcome of this perceptual normalization process is an abstract representation of letter identities that has been termed the visual word form (Riesenhuber and Poggio, 1999; Warrington and Shallice, 1980). We formulated the idea that an area in the midportion of the left fusiform gyrus, which activates wh

Despite the success of many pattern recognition systems in constrained domains, problems that involve noisy input and many classes remain difficult. A promising direction is to use several classifiers simultaneously, such that they can complement each other in correctness. This thesis is concerned with decision combination in a multiple classifier system that is critical to its success. A multiple classifier system consists of a set of classifiers and a decision combination function. It is a preferred solution to a complex recognition problem because it allows simultaneous use of feature descriptors of many types, corresponding measures of similarity, and many classification procedures. It also allows dynamic selection, so that classifiers adapted to inputs of a particular type may be applied only when those inputs are encountered. Decisions by the classifiers are represented as rankings of the class set that are derivable from the results of feature matching. Rank scores contain more ...

The effects of context on perceptual identification responses given without tinle pressure are well-described by classical models in which contextual and stimulus information exert independent effects. A recent Ilrticle by Massaro (1989) raises the possibility that interactive models, such as the TRACE model of speech perception, are inherently incompatible with these classical context effects. The present article shows that this incompatibility hypothesis can be rejected. Mathematical analysis and computer simulation methods are used to show that interactive models can exhibit the classical effects of context, if there is variability in the input to the network or if there is intrinsic variability in the network itself. A variety of interactive models which incorporate variability can aU produce the classical context effects, at least under some conditions; the conditions are rather general in the case of one of the variants. The findings suggest that interactive models should not be viewed as alternatives to classical accounts, but as hypotheses about the dynamics of information processing that lead to the global asymptotic behavior that the classical models describe.

similarity effects in masked associative priming One issue that all models of visual word recognition in alphabetic orthographies must ultimately take a position on is how the human processing system encodes letter positions when creating internal orthographic representations. Furthermore, although the choice of a coding scheme might seem to be a secondary aspect of these models, it can have a large impact on a model’s predictions (Andrews, 1996). For example, virtually all of the current models assume that the derived orthographic representation activates the lexical representations of formally similar words