Showing papers in "Neural Networks in 2008"

TL;DR: Research carried out on locomotor central pattern generators (CPGs), i.e. neural circuits capable of producing coordinated patterns of high-dimensional rhythmic output signals while receiving only simple, low-dimensional, input signals, is reviewed.

TL;DR: This paper examines learning of complex motor skills with human-like limbs, and combines the idea of modular motor control by means of motor primitives as a suitable way to generate parameterized control policies for reinforcement learning with the theory of stochastic policy gradient learning.

TL;DR: The results show that classifier performance deteriorates with even modest class imbalance in the training data and it is shown that BP is generally preferable over PSO for imbalanced training data especially with small data sample and large number of features.

TL;DR: A myoelectric-driven, finite state controller for a powered ankle-foot prosthesis that modulates both impedance and power output during stance is developed and evaluated, finding that the amputee can robustly transition between the finite state controllers through direct muscle activation, allowing rapid transitioning from level-ground to stair walking patterns.

TL;DR: The neuron cell provides an area and energy efficient implementation of the silicon cortical neuron, and could be used as a universal neuron circuit in VLSI neuromorphic networks that closely resemble the circuits of the cortex.

TL;DR: A Matlab/C toolbox, Brain-SMART (System for Multivariate AutoRegressive Time series, or BSMART), for spectral analysis of continuous neural time series data recorded simultaneously from multiple sensors designed for easy accessibility.

TL;DR: The p-delta rule is exhibited, which is a biologically more realistic alternative to backprop in biological neural circuits, but also for implementations in special purpose hardware and it is shown that its performance is competitive with that of other learning approaches from neural networks and machine learning.

TL;DR: This work constructs a biologically plausible hierarchy of neural detectors, which can discriminate seven basic emotional states from static views of associated body poses, and is evaluated against human test subjects on a recent set of stimuli manufactured for research on emotional body language.

TL;DR: Emergent strikes a balance between detailed, computationally expensive spiking neuron models and abstract, Bayesian or symbolic systems, which allows for the rapid development and successful execution of complex cognitive models while maintaining biological plausibility.

TL;DR: A new criterion of asymptotic stability is derived in terms of a linear matrix inequality (LMI), which can be efficiently solved via standard numerical software and proves to be less conservative than most of the existing results.

TL;DR: The paper provides a sound theoretical basis to Clifford neural computation by introducing the new concepts of isomorphic neurons and isomorphic representations and a unified training rule for Clifford MLPs.

TL;DR: The proposed Adjusted SOINn Classifier (ASC) is based on SOINN (self-organizing incremental neural network), it automatically learns the number of prototypes needed to determine the decision boundary, and learns new information without destroying old learned information.

TL;DR: In this paper, the authors present an application driven hardware exploration where they implement real-time, isolated digit speech recognition using a Liquid State Machine, a recurrent neural network of spiking neurons where only the output layer is trained.

TL;DR: This work implements the dynamical field model on a service robot and demonstrates how it learns 30 objects from a very small number of views (about 5 per object are sufficient), and illustrates how properties of feature binding emerge from this framework.

TL;DR: The structure of the FIND framework is outlined and its functionality, the measures of quality control, and the policies for developers and users are described, as well as two examples of complex analyses with FIND tools.

TL;DR: The Cerebellar Development Transcriptome Database not only provides a unique informatics tool for mining both spatial and temporal pattern information on gene expression in developing mouse brains, but also opens up opportunities to elucidate the transcriptome for cerebellar development.

TL;DR: The simulation of steady walking at 0.6 m/s of both the forelegs only and the hind legs only (with a supporting structure at the back and at the front respectively), achieved using the quadrupedal model is reported.

TL;DR: A boosting approach to random subspace method (RSM) to achieve an improved performance and avoid some of the major drawbacks of RSM, and can be used with any classifier, including those, such as k nearest neighbor classifiers, that cannot use boosting methods easily.

TL;DR: A multi-disciplinary approach at the convergence of neuroscience, dynamical system theory and autonomous robotics, in order to propose an efficient action selection mechanism based on a new model of the basal ganglia, which exhibits valuable dithering avoidance and energy-saving properties, when compared with a simple if-then-else decision rule.

TL;DR: A neural network controller for improved fuel efficiency of the Toyota Prius hybrid electric vehicle and a new method to detect and mitigate a battery fault is presented.

TL;DR: The experimental results show that the flexibility of the new, generalized, error function E(Exp), inspired by the Z-EDM algorithm, allows one to obtain the best results achievable with the other functions with a performance improvement in some cases.

TL;DR: A technique based on both Independent Component Analysis (ICA) to extract artifactual signals and on Renyi's entropy to automatically detect them is presented and is shown to be able to detect muscle and very low frequency activity as well as to discriminate them from other kinds of artifacts.

TL;DR: The results show that rule extraction with the GRG method produces rule sets that are more accurate and concise compared to those obtained by a decision tree method and an existing neural network rule extraction method.

TL;DR: A new machine learning paradigm intended for multi-class classification problems where the classes are ordered is introduced, namely when flexible discrete distributions, a new concept introduced here, are considered.

TL;DR: In this paper, a novel model employing the properties of thermodynamic systems operating far from equilibrium, which is analyzed by linearization near adaptive operating points using root locus techniques, is proposed.

TL;DR: The proposed integrated saliency map model includes an affective computing process that skips an unwanted area and pays attention to a desired area, which reflects the human preference and refusal in subsequent visual search processes.

TL;DR: In this article, the authors consider a discrimination task scenario to study language acquisition in which an agent receives linguistic input from an external teacher, in addition to sensory stimuli from the objects that exemplify the overlapping categories that make up the environment.

TL;DR: An adaptive neural network method is proposed to estimate two immeasurable physical parameters on-line and to compensate for the model uncertainty and engine time varying dynamics, so that the chattering is substantially reduced and the air-fuel ratio is regulated within the desired range of the stoichiometric value.

TL;DR: The SOVEREIGN animat model embodies these capabilities, and is tested in a 3D virtual reality environment, and can control different motor sequences under different motivational states and learns more efficient sequences to rewarded goals as exploration proceeds.