Results 1 
9 of
9
Learning to Perceive the World as Articulated: An Approach for Hierarchical Learning in SensoryMotor Systems
 NEURAL NETWORKS
, 1999
"... This paper describes how agents can learn an internal model of the world structurally by focusing on the problem of behaviorbased articulation. We develop an online learning scheme  the socalled mixture of recurrent neural net (RNN) experts  in which a set of RNN modules becomes selforgan ..."
Abstract

Cited by 94 (23 self)
 Add to MetaCart
This paper describes how agents can learn an internal model of the world structurally by focusing on the problem of behaviorbased articulation. We develop an online learning scheme  the socalled mixture of recurrent neural net (RNN) experts  in which a set of RNN modules becomes selforganized as experts on multiple levels in order to account for the different categories of sensorymotor flow which the robot experiences. Autonomous switching of activated modules in the lower level actually represents the articulation of the sensorymotor flow. In the meanwhile, a set of RNNs in the higher level competes to learn the sequences of module switching in the lower level, by which articulation at a further more abstract level can be achieved. The proposed scheme was examined through simulation experiments involving the navigation learning problem. Our dynamical systems analysis clarified the mechanism of the articulation; the possible correspondence between the articulation...
Modelbased Learning for Mobile Robot Navigation from the Dynamical Systems Perspective
 IEEE Transactions on Systems, Man, and Cybernetics
, 1996
"... This paper discusses how a behaviorbased robot can construct a “symbolic process” that accounts for its deliberative thinking processes using models of the environment. The paper focuses on two essential problems; one is the symbol grounding problem and the other is how the internal symbolic proces ..."
Abstract

Cited by 80 (20 self)
 Add to MetaCart
This paper discusses how a behaviorbased robot can construct a “symbolic process” that accounts for its deliberative thinking processes using models of the environment. The paper focuses on two essential problems; one is the symbol grounding problem and the other is how the internal symbolic processes can be situated with respect to the behavioral contexts. We investigate these problems by applying a dynamical system’s approach to the robot navigation learning problem. Our formulation, based on a forward modeling scheme using recurrent neural learning, shows that the robot is capable of learning grammatical structure hidden in the geometry of the workspace from the local sensory inputs through its navigational experiences. Furthermore, the robot is capable of generating diverse action plans to reach an arbitrary goal using the acquired forward model which incorporates chaotic dynamics. The essential claim is that the internal symbolic process, being embedded in the attractor, is grounded since it is selforganized solely through interaction with the physical world. It is also shown that structural stability arises in the interaction between the neural dynamics and the environmental dynamics, which accounts for the situatedness of the internal symbolic process. The experimental results using a mobile robot, equipped with a local sensor consisting of a laser range finder, verify our claims. 1 1
SelfOrganization of Modules and Their Hierarchy in Robot Learning Problems: A Dynamical Systems Approach
, 1997
"... This paper describes how the internal representation of the world can be selforganized in modular and hierarchical ways in a neural network architecture for sensorymotor systems. We develop an online learning scheme  the socalled mixture of recurrent neural net (RNN) experts  in which a set o ..."
Abstract

Cited by 8 (0 self)
 Add to MetaCart
This paper describes how the internal representation of the world can be selforganized in modular and hierarchical ways in a neural network architecture for sensorymotor systems. We develop an online learning scheme  the socalled mixture of recurrent neural net (RNN) experts  in which a set of RNN modules becomes selforganized as experts in multiple levels in order to account for the different categories of sensorymotor flow which the robot experiences. The proposed scheme was examined through simulation experiments involving the navigation learning problem, in which a robot equipped with range sensors traveled around rooms of different shape. It was shown that representative building blocks or "concepts" corresponding to turning right and left at corners, going straight along corridors and encountering junctions are selforganized in their respective modules in the lower level network. In the higher level network, the "concepts" corresponding to traveling in different rooms a...
Bifurcations of Periodic Solutions in a Coupled Oscillator with Voltage Ports
, 1998
"... this paper, we study bifurcations of equilibrium points and periodic solutions observed in a resistively coupled oscillator with voltage ports. We classify equilibrium points and periodic solutions into four and eight different types, respectively, according to their symmetrical properties. By calcu ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
this paper, we study bifurcations of equilibrium points and periodic solutions observed in a resistively coupled oscillator with voltage ports. We classify equilibrium points and periodic solutions into four and eight different types, respectively, according to their symmetrical properties. By calculating Dtype of branching sets (symmetrybreaking bifurcations) of equilibrium points and periodic solutions, we show that all types of equilibrium points and periodic solutions are systematically found. Possible oscillations in two coupled oscillators are presented by calculating Hopf bifurcation sets of equilibrium points. A parameter region in which chaotic oscillations exist is also shown by obtaining a cascade of perioddoubling bifurcation sets
Synchronization of four coupled van der Pol oscillators
, 2008
"... It is possible that selfexcited vibrations in turbine blades synchronize due to elastic coupling through the shaft. In this context, the synchronization of four coupled van der Pol oscillators is studied here. For quasilinear oscillations, a stability condition is derived from an analysis based on ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
It is possible that selfexcited vibrations in turbine blades synchronize due to elastic coupling through the shaft. In this context, the synchronization of four coupled van der Pol oscillators is studied here. For quasilinear oscillations, a stability condition is derived from an analysis based on linearizing the original equation around an unperturbed limit cycle and transforming it into Hill’s equation. For the nonlinear case, numerical simulations show the existence of two welldefined regions of phase relationships in parameter space in which a multiplicity of periodic attractors is embedded. The size of these regions strongly depends on the values of the uncoupled oscillator and coupling constants. For the coupling constant below a critical value, there exists a region in which a diversity of phaseshift attractors is present, whereas for values above the critical an inphase attractor is predominant. It was observed that the presence of an antiphase attractor in the subcritical region is associated with sudden changes in the period of the coupled oscillators. The convergence of the coupled system to a particular periodic attractor is explored using several initial conditions from all zones of the parameter space. The study was extended to nonidentical oscillators, and it was found that there is synchronization even over a wide range of dissimilarity. 1
Mutual Synchronization In 4 Coupled Oscillators With Different Natural Frequencies
, 1998
"... In this paper, we investigate the synchronization phenomena in 4 oscillators with different natural frequencies fullcoupled by capacitors. When there is no natural frequency difference among the oscillators, 4phase oscillation is not stable. In this system, however, because the frequencies of the ..."
Abstract
 Add to MetaCart
In this paper, we investigate the synchronization phenomena in 4 oscillators with different natural frequencies fullcoupled by capacitors. When there is no natural frequency difference among the oscillators, 4phase oscillation is not stable. In this system, however, because the frequencies of the oscillators have a slight difference, nearly sinusoidal 4phase oscillations are stably excited. Moreover we have proved the stability of 4phase solutions by calculating the stability of the fixed points and considered why frequency locking occurs. From the results we can confirm that 4phase oscillations can be excited when the symmetry of the system collapses. 1. INTRODUCTION Mutual synchronization in the coupled oscillatory systems is a very important feature in nonlinear science and many researches have been carried out [1][5]. Recently, in particular, we have investigated the system in which 4 or more oscillators are coupled by one resistor, and reported that N  phase oscillation...
Spectral Analysis and Dynamical Behavior of Complex Networks
"... Abstract—In this paper, we employ spectral graph theory as a tool for analyzing the Internet topology. We show its importance in understanding dynamical behavior of complex networks. We also provide an overview of various approaches dealing with synchronization in complex networks. 1. ..."
Abstract
 Add to MetaCart
Abstract—In this paper, we employ spectral graph theory as a tool for analyzing the Internet topology. We show its importance in understanding dynamical behavior of complex networks. We also provide an overview of various approaches dealing with synchronization in complex networks. 1.
Hierarchical Learning in SensoryMotor Systems
"... This paper describes how agents can learn an internal model of the world structurally by focusing on the problem of behaviorbased articulation. We develop an online learning scheme { the socalled mixture of recurrent neural net (RNN) experts { in which a set of RNN modules becomes selforganized ..."
Abstract
 Add to MetaCart
This paper describes how agents can learn an internal model of the world structurally by focusing on the problem of behaviorbased articulation. We develop an online learning scheme { the socalled mixture of recurrent neural net (RNN) experts { in which a set of RNN modules becomes selforganized as experts on multiple levels in order to account for the di erent categories of sensorymotor ow which the robot experiences. Autonomous switching of activated modules in the lower level actually represents the articulation of the sensorymotor ow. In the meanwhile, a set of RNNs in the higher level competes to learn the sequences of module switching in the lower level, by which articulation at a further more abstract level can be achieved. The proposed scheme was examined through simulation experiments involving the navigation learning problem. Our dynamical systems analysis clari ed the mechanism of the articulation; the possible correspondence between the articulation mechanism and the attention switching mechanism in thalamocortical loops is discussed.