A NEURAL NET FOR EXTRACTING KNOWLEDGE FROM NATURAL-LANGUAGE DATA-BASES

The present paper introduces a new model of fuzzy neuron, one which increases the computational power of the artificial neuron, turning it also into a symbolic processing device. This model proposes the synapsis to be symbolically and numerically defined, by means of the assignment of tokens to the presynaptic and postsynaptic neurons. The matching or concatenation compatibility between these tokens is used to decided about the possible connections among neurons of a given net. The strength of the compatible synapsis is made dependent on the amount of the available presynaptic and post synaptic tokens. The symbolic and numeric processing capacity of the new fuzzy neuron is used here to build a neural net (JARGON) to disclose the existing knowledge in natural language data bases such as medical files, set of interviews, and reports about engineering operations.

Analysis of Structural Integrity Using an ARTMAP-Fuzzy Artificial Neural Network

This paper presents the application of artificial neural networks in the analysis of the structural integrity of a building. The main objective is to apply an artificial neural network based on adaptive resonance theory, called ARTMAP-Fuzzy neural network and apply it to the identification and characterization of structural failure. This methodology can help professionals in the inspection of structures, to identify and characterize flaws in order to conduct preventative maintenance to ensure the integrity of the structure and decision-making. In order to validate the methodology was modeled a building of two walk, and from this model were simulated various situations (base-line condition and improper conditions), resulting in a database of signs, which were used as input data for ARTMAP-Fuzzy network. The results show efficiency, robustness and accuracy.

Artificial neural network associated to UV/Vis spectroscopy for monitoring bioreactions in biopharmaceutical processes

Currently, mammalian cells are the most utilized hosts for biopharmaceutical production. The culture media for these cell lines include commonly in their composition a pH indicator. Spectroscopic techniques are used for biopharmaceutical process monitoring, among them, UV–Vis spectroscopy has found scarce applications. This work aimed to define artificial neural networks architecture and fit its parameters to predict some nutrients and metabolites, as well as viable cell concentration based on UV–Vis spectral data of mammalian cell bioprocess using phenol red in culture medium. The BHK-21 cell line was used as a mammalian cell model. Off-line spectra of supernatant samples taken from batches performed at different dissolved oxygen concentrations in two bioreactor configurations and with two pH control strategies were used to define two artificial neural networks. According to absolute errors, glutamine (0.13 ± 0.14 mM), glutamate (0.02 ± 0.02 mM), glucose (1.11 ± 1.70 mM), lactate (0.84 ± 0.68 mM) and viable cell concentrations (1.89 105 ± 1.90 105 cell/mL) were suitably predicted. The prediction error averages for monitored variables were lower than those previously reported using different spectroscopic techniques in combination with partial least squares or artificial neural network. The present work allows for UV–VIS sensor development, and decreases cost related to nutrients and metabolite quantifications.

A structural model of emotions of cognitive dissonances

Cognitive dissonance is the stress that comes from holding two conflicting thoughts simultaneously in the mind, usually arising when people are asked to choose between two detrimental or two beneficial options. In view of the well-established role of emotions in decision making, here we investigate whether the conventional structural models used to represent the relationships among basic emotions, such as the Circumplex model of affect, can describe the emotions of cognitive dissonance as well. We presented a questionnaire to 34 anonymous participants, where each question described a decision to be made among two conflicting motivations and asked the participants to rate analogically the pleasantness and the intensity of the experienced emotion. We found that the results were compatible with the predictions of the Circumplex model for basic emotions. (C) 2012 Elsevier Ltd. All rights reserved.

Phenol biodegradation by a microbial consortium: application of artificial neural network (ANN) modelling

In this study, an effective microbial consortium for the biodegradation of phenol was grown under different operational conditions, and the effects of phosphate concentration (1.4 g L-1, 2.8 g L-1, 4.2 g L-1), temperature (25 degrees C, 30 degrees C, 35 degrees C), agitation (150 rpm, 200 rpm, 250 rpm) and pH (6, 7, 8) on phenol degradation were investigated, whereupon an artificial neural network (ANN) model was developed in order to predict degradation. The learning, recall and generalization characteristics of neural networks were studied using data from the phenol degradation system. The efficiency of the model generated by the ANN was then tested and compared with the experimental results obtained. In both cases, the results corroborate the idea that aeration and temperature are crucial to increasing the efficiency of biodegradation.

A decaying factor accounts for contained activity in neuronal networks with no need of hierarchical or modular organization

The mechanisms responsible for containing activity in systems represented by networks are crucial in various phenomena, for example, in diseases such as epilepsy that affect the neuronal networks and for information dissemination in social networks. The first models to account for contained activity included triggering and inhibition processes, but they cannot be applied to social networks where inhibition is clearly absent. A recent model showed that contained activity can be achieved with no need of inhibition processes provided that the network is subdivided into modules (communities). In this paper, we introduce a new concept inspired in the Hebbian theory, through which containment of activity is achieved by incorporating a dynamics based on a decaying activity in a random walk mechanism preferential to the node activity. Upon selecting the decay coefficient within a proper range, we observed sustained activity in all the networks tested, namely, random, Barabasi-Albert and geographical networks. The generality of this finding was confirmed by showing that modularity is no longer needed if the dynamics based on the integrate-and-fire dynamics incorporated the decay factor. Taken together, these results provide a proof of principle that persistent, restrained network activation might occur in the absence of any particular topological structure. This may be the reason why neuronal activity does not spread out to the entire neuronal network, even when no special topological organization exists.

Stochastic competitive learning in complex networks

Competitive learning is an important machine learning approach which is widely employed in artificial neural networks. In this paper, we present a rigorous definition of a new type of competitive learning scheme realized on large-scale networks. The model consists of several particles walking within the network and competing with each other to occupy as many nodes as possible, while attempting to reject intruder particles. The particle's walking rule is composed of a stochastic combination of random and preferential movements. The model has been applied to solve community detection and data clustering problems. Computer simulations reveal that the proposed technique presents high precision of community and cluster detections, as well as low computational complexity. Moreover, we have developed an efficient method for estimating the most likely number of clusters by using an evaluator index that monitors the information generated by the competition process itself. We hope this paper will provide an alternative way to the study of competitive learning.

Phase-disorder-induced firing activity in excitable neuronal networks with attractive and repulsive coupling

It has been revealed that the network of excitable neurons via attractive coupling can generate spikes under stimuli of subthreshold signals with disordered phases. In this paper, we explore the firing activity induced by phase disorder in excitable neuronal networks consisting of both attractive and repulsive coupling. By increasing the fraction of repulsive coupling, we find that, in the weak coupling strength case, the firing threshold of phase disorder is increased and the system response to subthreshold signals is decreased, indicating that the effect of inducing neuron firing by phase disorder is weakened with repulsive coupling. Interestingly, in the large coupling strength case, we see an opposite situation, where the coupled neurons show a rather large response to the subthreshold signals even with small phase disorder. The latter case implies that the effect of phase disorder is enhanced by repulsive coupling. A system of two-coupled excitable neurons is used to explain the role of repulsive coupling on phase-disorder-induced firing activity.

Simulating event related potentials in tasks involving conflict resolution: a neural network

In this thesis, the main Executive Control theories are exposed. Methods typical of Cognitive and Computational Neuroscience are introduced and the role of behavioural tasks involving conflict resolution in the response elaboration, after the presentation of a stimulus to the subject, are highlighted. In particular, the Eriksen Flanker Task and its variants are discussed. Behavioural data, from scientific literature, are illustrated in terms of response times and error rates. During experimental behavioural tasks, EEG is registered simultaneously. Thanks to this, event related potential, related with the current task, can be studied. Different theories regarding relevant event related potential in this field - such as N2, fERN (feedback Error Related Negativity) and ERN (Error Related Negativity) – are introduced. The aim of this thesis is to understand and simulate processes regarding Executive Control, including performance improvement, error detection mechanisms, post error adjustments and the role of selective attention, with the help of an original neural network model. The network described here has been built with the purpose to simulate behavioural results of a four choice Eriksen Flanker Task. Model results show that the neural network can simulate response times, error rates and event related potentials quite well. Finally, results are compared with behavioural data and discussed in light of the mentioned Executive Control theories. Future perspective for this new model are outlined.

Fluctuation properties in random walks on networks and simple integrate and fire models

In questa tesi si è studiato l’insorgere di eventi critici in un semplice modello neurale del tipo Integrate and Fire, basato su processi dinamici stocastici markoviani definiti su una rete. Il segnale neurale elettrico è stato modellato da un flusso di particelle. Si è concentrata l’attenzione sulla fase transiente del sistema, cercando di identificare fenomeni simili alla sincronizzazione neurale, la quale può essere considerata un evento critico. Sono state studiate reti particolarmente semplici, trovando che il modello proposto ha la capacità di produrre effetti "a cascata" nell’attività neurale, dovuti a Self Organized Criticality (auto organizzazione del sistema in stati instabili); questi effetti non vengono invece osservati in Random Walks sulle stesse reti. Si è visto che un piccolo stimolo random è capace di generare nell’attività della rete delle fluttuazioni notevoli, in particolar modo se il sistema si trova in una fase al limite dell’equilibrio. I picchi di attività così rilevati sono stati interpretati come valanghe di segnale neurale, fenomeno riconducibile alla sincronizzazione.

Development of Dynamic Constraint Models for a Model Based Transient Calibration Process

Model based calibration has gained popularity in recent years as a method to optimize increasingly complex engine systems. However virtually all model based techniques are applied to steady state calibration. Transient calibration is by and large an emerging technology. An important piece of any transient calibration process is the ability to constrain the optimizer to treat the problem as a dynamic one and not as a quasi-static process. The optimized air-handling parameters corresponding to any instant of time must be achievable in a transient sense; this in turn depends on the trajectory of the same parameters over previous time instances. In this work dynamic constraint models have been proposed to translate commanded to actually achieved air-handling parameters. These models enable the optimization to be realistic in a transient sense. The air handling system has been treated as a linear second order system with PD control. Parameters for this second order system have been extracted from real transient data. The model has been shown to be the best choice relative to a list of appropriate candidates such as neural networks and first order models. The selected second order model was used in conjunction with transient emission models to predict emissions over the FTP cycle. It has been shown that emission predictions based on air-handing parameters predicted by the dynamic constraint model do not differ significantly from corresponding emissions based on measured air-handling parameters.

Development of a model-based transient calibration process for diesel engine electronic control module tables – Part 2: modelling and optimization

This is the second part of a study investigating a model-based transient calibration process for diesel engines. The first part addressed the data requirements and data processing required for empirical transient emission and torque models. The current work focuses on modelling and optimization. The unexpected result of this investigation is that when trained on transient data, simple regression models perform better than more powerful methods such as neural networks or localized regression. This result has been attributed to extrapolation over data that have estimated rather than measured transient air-handling parameters. The challenges of detecting and preventing extrapolation using statistical methods that work well with steady-state data have been explained. The concept of constraining the distribution of statistical leverage relative to the distribution of the starting solution to prevent extrapolation during the optimization process has been proposed and demonstrated. Separate from the issue of extrapolation is preventing the search from being quasi-static. Second-order linear dynamic constraint models have been proposed to prevent the search from returning solutions that are feasible if each point were run at steady state, but which are unrealistic in a transient sense. Dynamic constraint models translate commanded parameters to actually achieved parameters that then feed into the transient emission and torque models. Combined model inaccuracies have been used to adjust the optimized solutions. To frame the optimization problem within reasonable dimensionality, the coefficients of commanded surfaces that approximate engine tables are adjusted during search iterations, each of which involves simulating the entire transient cycle. The resulting strategy, different from the corresponding manual calibration strategy and resulting in lower emissions and efficiency, is intended to improve rather than replace the manual calibration process.

Dynamics of Spatial Game Theory Networks with Novel Modifications

Introduction: Advances in biotechnology have shed light on many biological processes. In biological networks, nodes are used to represent the function of individual entities within a system and have historically been studied in isolation. Network structure adds edges that enable communication between nodes. An emerging fieldis to combine node function and network structure to yield network function. One of the most complex networks known in biology is the neural network within the brain. Modeling neural function will require an understanding of networks, dynamics, andneurophysiology. It is with this work that modeling techniques will be developed to work at this complex intersection. Methods: Spatial game theory was developed by Nowak in the context of modeling evolutionary dynamics, or the way in which species evolve over time. Spatial game theory offers a two dimensional view of analyzingthe state of neighbors and updating based on the surroundings. Our work builds upon this foundation by studying evolutionary game theory networks with respect to neural networks. This novel concept is that neurons may adopt a particular strategy that will allow propagation of information. The strategy may therefore act as the mechanism for gating. Furthermore, the strategy of a neuron, as in a real brain, isimpacted by the strategy of its neighbors. The techniques of spatial game theory already established by Nowak are repeated to explain two basic cases and validate the implementation of code. Two novel modifications are introduced in Chapters 3 and 4 that build on this network and may reflect neural networks. Results: The introduction of two novel modifications, mutation and rewiring, in large parametricstudies resulted in dynamics that had an intermediate amount of nodes firing at any given time. Further, even small mutation rates result in different dynamics more representative of the ideal state hypothesized. Conclusions: In both modificationsto Nowak's model, the results demonstrate the network does not become locked into a particular global state of passing all information or blocking all information. It is hypothesized that normal brain function occurs within this intermediate range and that a number of diseases are the result of moving outside of this range.

Advantages and Applications of Transforming Empirical Model Input Space with Dimensional Models

Model-based calibration of steady-state engine operation is commonly performed with highly parameterized empirical models that are accurate but not very robust, particularly when predicting highly nonlinear responses such as diesel smoke emissions. To address this problem, and to boost the accuracy of more robust non-parametric methods to the same level, GT-Power was used to transform the empirical model input space into multiple input spaces that simplified the input-output relationship and improved the accuracy and robustness of smoke predictions made by three commonly used empirical modeling methods: Multivariate Regression, Neural Networks and the k-Nearest Neighbor method. The availability of multiple input spaces allowed the development of two committee techniques: a 'Simple Committee' technique that used averaged predictions from a set of 10 pre-selected input spaces chosen by the training data and the "Minimum Variance Committee" technique where the input spaces for each prediction were chosen on the basis of disagreement between the three modeling methods. This latter technique equalized the performance of the three modeling methods. The successively increasing improvements resulting from the use of a single best transformed input space (Best Combination Technique), Simple Committee Technique and Minimum Variance Committee Technique were verified with hypothesis testing. The transformed input spaces were also shown to improve outlier detection and to improve k-Nearest Neighbor performance when predicting dynamic emissions with steady-state training data. An unexpected finding was that the benefits of input space transformation were unaffected by changes in the hardware or the calibration of the underlying GT-Power model.

Learning real-world stimuli in a neural network with spike-driven synaptic dynamics

We present a model of spike-driven synaptic plasticity inspired by experimental observations and motivated by the desire to build an electronic hardware device that can learn to classify complex stimuli in a semisupervised fashion. During training, patterns of activity are sequentially imposed on the input neurons, and an additional instructor signal drives the output neurons toward the desired activity. The network is made of integrate-and-fire neurons with constant leak and a floor. The synapses are bistable, and they are modified by the arrival of presynaptic spikes. The sign of the change is determined by both the depolarization and the state of a variable that integrates the postsynaptic action potentials. Following the training phase, the instructor signal is removed, and the output neurons are driven purely by the activity of the input neurons weighted by the plastic synapses. In the absence of stimulation, the synapses preserve their internal state indefinitely. Memories are also very robust to the disruptive action of spontaneous activity. A network of 2000 input neurons is shown to be able to classify correctly a large number (thousands) of highly overlapping patterns (300 classes of preprocessed Latex characters, 30 patterns per class, and a subset of the NIST characters data set) and to generalize with performances that are better than or comparable to those of artificial neural networks. Finally we show that the synaptic dynamics is compatible with many of the experimental observations on the induction of long-term modifications (spike-timing-dependent plasticity and its dependence on both the postsynaptic depolarization and the frequency of pre- and postsynaptic neurons).