MODULATED EXPLORATORY DYNAMICS CAN SHAPE SELF-ORGANIZED BEHAVIOR

2009 ◽  
Vol 12 (03) ◽  
pp. 273-291 ◽  
Author(s):  
FRANK HESSE ◽  
RALF DER ◽  
J. MICHAEL HERRMANN
Keyword(s):  
Adaptive Controller ◽  
Second Order ◽  
Learning Mechanism ◽  
Low Level ◽  
Context Sensitive ◽  
Self Organized ◽  
Learning Rules ◽  
Internal Parameters ◽  
Source Of Information ◽  
Humanoid Hand

We study an adaptive controller that adjusts its internal parameters by self-organization of its interaction with the environment. We show that the parameter changes that occur in this low-level learning process can themselves provide a source of information to a higher-level context-sensitive learning mechanism. In this way, the context is interpreted in terms of the concurrent low-level learning mechanism. The dual learning architecture is studied in realistic simulations of a foraging robot and of a humanoid hand that manipulated an object. Both systems are driven by the same low-level scheme, but use the second-order information in different ways. While the low-level adaptation continues to follow a set of rigid learning rules, the second-order learning modulates the elementary behaviors and affects the distribution of the sensory inputs via the environment.

Fast Learning in Networks of Locally-Tuned Processing Units

1989 ◽  
Vol 1 (2) ◽  
pp. 281-294 ◽  
Author(s):  
John Moody ◽  
Christian J. Darken
Keyword(s):  
Network Architecture ◽  
Hybrid Learning ◽  
Computationally Efficient ◽  
Fast Learning ◽  
Computational Overhead ◽  
Self Organized ◽  
Learning Rules ◽  
Classification Tasks ◽  
Function Approximations ◽  
Local Representations

We propose a network architecture which uses a single internal layer of locally-tuned processing units to learn both classification tasks and real-valued function approximations (Moody and Darken 1988). We consider training such networks in a completely supervised manner, but abandon this approach in favor of a more computationally efficient hybrid learning method which combines self-organized and supervised learning. Our networks learn faster than backpropagation for two reasons: the local representations ensure that only a few units respond to any given input, thus reducing computational overhead, and the hybrid learning rules are linear rather than nonlinear, thus leading to faster convergence. Unlike many existing methods for data analysis, our network architecture and learning rules are truly adaptive and are thus appropriate for real-time use.

Self-Organization Processes of Nanostructure Formation by Laser Radiation in Metal

2011 ◽  
Vol 222 ◽  
pp. 18-23 ◽  
Author(s):  
L. Fedorenko
Keyword(s):  
Plasmon Resonance ◽  
Active Zone ◽  
Rarefied Gas ◽  
High Energy ◽  
Nano Particles ◽  
Self Organization ◽  
Self Organized ◽  
Current State ◽  
Internal Parameters ◽  
Metal Nano Particles

The current state of knowledge about mechanisms of metal nano-particles (NP) formation processes induced by the interaction of high-energy laser beam with surface of the metallic lattices (Au, Ag, Cu) is presented. The review includes an evaluation of the contribution of self-organized effects into the processes of the metal nano-structurization depending on the laser mode, external factors and internal parameters of an active zone. It was noticed that intensive pulsed laser illumination enabled to stimulate nano-fragmentation at the fluencies near and above the melting threshold of the metal in different mediums The laser induced processes of metallic particles formation by ablation of the metal target with consequent NP sizes stabilization by precise temperature tuning in the active zone, local plasmon resonance in liquids, and microablation mechanism in metal films in the conditions of surface plasmon resonance (SPR) due to self-organization effects are considered. Comparative analysis of the laser nano-technologies in air, vacuum, rarefied gas and liquid environments showed the advantages of self-organization in NP generation processes based on the SPR effects and their perspectives.

scholarly journals RATING OF USERS INVOLVEMENT IN THE RADICAL GROUPS IN SOME SOCIAL NETWORKS

2019 ◽  
pp. 113-122
Author(s):  
N.S. Barabash ◽  
D.S. Zhukov
Keyword(s):  
Social Networks ◽  
Hong Kong ◽  
Pink Noise ◽  
Self Organized ◽  
Self Organized Criticality ◽  
High Level ◽  
Facebook Pages ◽  
Source Of Information

This issue is an enclosure to the theory of self-organized criticality (SOC) for studying of the radical protest groups of people in some social networks. The SOC theory needs for the value of level of the users involvement in Facebook communities which support the protest moves in Hong Kong in 2019. There were studied about 35 Facebook pages. The period of studying 01.03 to 23.03.2019. This article claims that the communities with the high level of users involvement are based on self-organized criticality. This item also explains some SOC theory approaches according to which a method of pink noise identity is one of the SOC attribute. It is necessary to say that some protest communities work in SOC regime. In spite of the seeming polycentric of the protest network the connection of reflection comes to a few number of Facebook pages which are the source of information and intension of the protesters so they can become the event’s drivers.

scholarly journals Self-Learning Low-Level Controllers

2021 ◽  
Author(s):  
Dang Xuan Ba ◽  
Joonbum Bae
Keyword(s):  
Sliding Mode ◽  
Humanoid Robots ◽  
Nonlinear Functions ◽  
Low Level ◽  
Unstructured Environments ◽  
Learning Rules ◽  
Whole Systems ◽  
Self Learning ◽  
Control Criterion ◽  
Adaptation Ability

Humanoid robots are complicated systems both in hardware and software designs. Furthermore, the robots normally work in unstructured environments at which unpredictable disturbances could degrade control performances of whole systems. As a result, simple yet effective controllers are favorite employed in low-level layers. Gain-learning algorithms applied to conventional control frameworks, such as Proportional-Integral-Derivative, Sliding-mode, and Backstepping controllers, could be reasonable solutions. The adaptation ability integrated is adopted to automatically tune proper control gains subject to the optimal control criterion both in transient and steady-state phases. The learning rules could be realized by using analytical nonlinear functions. Their effectiveness and feasibility are carefully discussed by theoretical proofs and experimental discussion.

Visual Awareness and Visual Appearances

2018 ◽  
Author(s):  
Jérôme Dokic
Keyword(s):  
Visual Experience ◽  
Natural Kinds ◽  
Visual Awareness ◽  
Analytic Philosophy ◽  
Low Level ◽  
Context Sensitive ◽  
General Account ◽  
Volumetric Objects

Many philosophers from the traditions of both phenomenology and analytic philosophy have observed that our perceptual (e.g. visual) experience involves a certain duality. In the terminology used in this chapter, we seem to be visually aware of more than what is visually apparent to us. Such duality is present in various cases, from the perception of opaque volumetric objects to that of natural kinds, artefacts, and familiar persons. This chapter offers a general account of the duality, according to which visual appearances supervene on low-level visual facts while the scope of visual awareness depends on context-sensitive cognitive habits or heuristics based on visual appearances.

Sign in / Sign up

Export Citation Format

Share Document