arXiv subjects

Carlos Gershenson

Publications and source records attributed to Carlos Gershenson.

At least 109 records · Page 6Linked to original sources

arXiv2004

The role of behavior modifiers in representation development

We address the problem of the development of representations and their relationship to the environment. We study a software agent which develops in a network a representation of its simple environment which captures and integrates the relationships between agent and environment through a closure mechanism. The inclusion of a variable behavior modifier allows better representation development. This can be confirmed with an internal description of the closure mechanism, and with an external description of the properties of the representation network.

cs.AI↗

arXiv2003

Phase Transitions in Random Boolean Networks with Different Updating Schemes

In this paper we study the phase transitions of different types of Random Boolean networks. These differ in their updating scheme: synchronous, semi-synchronous, or asynchronous, and deterministic or non-deterministic. It has been shown that the statistical properties of Random Boolean networks change considerable according to the updating scheme. We study with computer simulations sensitivity to initial conditions as a measure of order/chaos. We find that independently of their updating scheme, all network types have very similar phase transitions, namely when the average number of connections of nodes is between one and three. This critical value depends more on the size of the network than on the updating scheme.

nlin.AO↗

arXiv2003

Self-organizing Traffic Control: First Results

We developed a virtual laboratory for traffic control where agents use different strategies in order to self-organize on the road. We present our first results where we compare the performance and behaviour promoted by environmental constrains and five different simple strategies: three inspired in flocking behaviour, one selfish, and one inspired in the minority game. Experiments are presented for comparing the strategies. Different issues are discussed, such as the important role of environmental constrains and the emergence of traffic lanes.

nlin.AO↗

arXiv2003

Artificial Neural Networks for Beginners

The scope of this teaching package is to make a brief induction to Artificial Neural Networks (ANNs) for people who have no previous knowledge of them. We first make a brief introduction to models of networks, for then describing in general terms ANNs. As an application, we explain the backpropagation algorithm, since it is widely used and many other algorithms are derived from it. The user should know algebra and the handling of functions and vectors. Differential calculus is recommendable, but not necessary. The contents of this package should be understood by people with high school education. It would be useful for people who are just curious about what are ANNs, or for people who want to become familiar with them, so when they study them more fully, they will already have clear notions of ANNs. Also, people who only want to apply the backpropagation algorithm without a detailed and formal explanation of it will find this material useful. This work should not be seen as "Nets for dummies", but of course it is not a treatise. Much of the formality is skipped for the sake of simplicity. Detailed explanations and demonstrations can be found in the referred readings. The included exercises complement the understanding of the theory. The on-line resources are highly recommended for extending this brief induction.

cs.NE↗

arXiv2003

Time-scales, Meaning, and Availability of Information in a Global Brain

We note the importance of time-scales, meaning, and availability of information for the emergence of novel information meta-structures at a global scale. We discuss previous work in this area and develop future perspectives. We focus on the transmission of scientific articles and the integration of traditional conferences with their virtual extensions on the Internet, their time-scales, and availability. We mention the Semantic Web as an effort for integrating meaningful information.

cs.AI↗

arXiv2003

When Can we Call a System Self-organizing?

We do not attempt to provide yet another definition of selforganization, but explore the conditions under which we can model a system as self-organizing. These involve the dynamics of entropy, and the purpose, aspects, and description level chosen by an observer. We show how, changing the level or "graining" of description, the same system can appear selforganizing or self-disorganizing. We discuss ontological issues we face when studying self-organizing systems, and analyse when designing and controlling artificial self-organizing systems is useful. We conclude that self-organization is a way of observing systems, not an absolute class of systems.

nlin.AO↗

arXiv2003

Contextual Random Boolean Networks

We propose the use of Deterministic Generalized Asynchronous Random Boolean Networks [Gershenson, 2002] as models of contextual deterministic discrete dynamical systems. We show that changes in the context have drastic effects on the global properties of the same networks, namely the average number of attractors and the average percentage of states in attractors. We introduce the situation where we lack knowledge on the context as a more realistic model for contextual dynamical systems. We notice that this makes the network non-deterministic in a specific way, namely introducing a non-Kolmogorovian quantum-like structure for the modelling of the network [Aerts, 1986]. In this case, for example, a state of the network has the potentiality (probability) of collapsing into different attractors, depending on the specific form of lack of knowledge on the context.

nlin.AO↗

arXiv2003

On the Notion of Cognition

We discuss philosophical issues concerning the notion of cognition basing ourselves in experimental results in cognitive sciences, especially in computer simulations of cognitive systems. There have been debates on the "proper" approach for studying cognition, but we have realized that all approaches can be in theory equivalent. Different approaches model different properties of cognitive systems from different perspectives, so we can only learn from all of them. We also integrate ideas from several perspectives for enhancing the notion of cognition, such that it can contain other definitions of cognition as special cases. This allows us to propose a simple classification of different types of cognition.

cs.AI↗

arXiv2002

Dynamic Adjustment of the Motivation Degree in an Action Selection Mechanism

This paper presents a model for dynamic adjustment of the motivation degree, using a reinforcement learning approach, in an action selection mechanism previously developed by the authors. The learning takes place in the modification of a parameter of the model of combination of internal and external stimuli. Experiments that show the claimed properties are presented, using a VR simulation developed for such purposes. The importance of adaptation by learning in action selection is also discussed.

cs.AI↗

arXiv2002

Modelling intracellular signalling networks using behaviour-based systems and the blackboard architecture

This paper proposes to model the intracellular signalling networks using a fusion of behaviour-based systems and the blackboard architecture. In virtue of this fusion, the model developed by us, which has been named Cellulat, allows to take account two essential aspects of the intracellular signalling networks: (1) the cognitive capabilities of certain types of networks' components and (2) the high level of spatial organization of these networks. A simple example of modelling of Ca2+ signalling pathways using Cellulat is presented here. An intracellular signalling virtual laboratory is being developed from Cellulat.

cs.MA↗

arXiv2002

Integration of Computational Techniques for the Modelling of Signal Transduction

A cell can be seen as an adaptive autonomous agent or as a society of adaptive autonomous agents, where each can exhibit a particular behaviour depending on its cognitive capabilities. We present an intracellular signalling model obtained by integrating several computational techniques into an agent-based paradigm. Cellulat, the model, takes into account two essential aspects of the intracellular signalling networks: cognitive capacities and a spatial organization. Exemplifying the functionality of the system by modelling the EGFR signalling pathway, we discuss the methodology as well as the purposes of an intracellular signalling virtual laboratory, presently under development.

cs.MA↗

arXiv2002

Adaptive Development of Koncepts in Virtual Animats: Insights into the Development of Knowledge

As a part of our effort for studying the evolution and development of cognition, we present results derived from synthetic experimentations in a virtual laboratory where animats develop koncepts adaptively and ground their meaning through action. We introduce the term "koncept" to avoid confusions and ambiguity derived from the wide use of the word "concept". We present the models which our animats use for abstracting koncepts from perceptions, plastically adapt koncepts, and associate koncepts with actions. On a more philosophical vein, we suggest that knowledge is a property of a cognitive system, not an element, and therefore observer-dependent.

cs.AI↗

arXiv2002

Thinking Adaptive: Towards a Behaviours Virtual Laboratory

In this paper we name some of the advantages of virtual laboratories; and propose that a Behaviours Virtual Laboratory should be useful for both biologists and AI researchers, offering a new perspective for understanding adaptive behaviour. We present our development of a Behaviours Virtual Laboratory, which at this stage is focused in action selection, and show some experiments to illustrate the properties of our proposal, which can be accessed via Internet.

cs.AI↗

arXiv2002

A Comparison of Different Cognitive Paradigms Using Simple Animats in a Virtual Laboratory, with Implications to the Notion of Cognition

In this thesis I present a virtual laboratory which implements five different models for controlling animats: a rule-based system, a behaviour-based system, a concept-based system, a neural network, and a Braitenberg architecture. Through different experiments, I compare the performance of the models and conclude that there is no "best" model, since different models are better for different things in different contexts. The models I chose, although quite simple, represent different approaches for studying cognition. Using the results as an empirical philosophical aid, I note that there is no "best" approach for studying cognition, since different approaches have all advantages and disadvantages, because they study different aspects of cognition from different contexts. This has implications for current debates on "proper" approaches for cognition: all approaches are a bit proper, but none will be "proper enough". I draw remarks on the notion of cognition abstracting from all the approaches used to study it, and propose a simple classification for different types of cognition.

cs.AI↗

arXiv2002

Classification of Random Boolean Networks

We provide the first classification of different types of Random Boolean Networks (RBNs). We study the differences of RBNs depending on the degree of synchronicity and determinism of their updating scheme. For doing so, we first define three new types of RBNs. We note some similarities and differences between different types of RBNs with the aid of a public software laboratory we developed. Particularly, we find that the point attractors are independent of the updating scheme, and that RBNs are more different depending on their determinism or non-determinism rather than depending on their synchronicity or asynchronicity. We also show a way of mapping non-synchronous deterministic RBNs into synchronous RBNs. Our results are important for justifying the use of specific types of RBNs for modelling natural phenomena.

cs.CC↗

arXiv2002

Behaviour-based Knowledge Systems: An Epigenetic Path from Behaviour to Knowledge

In this paper we expose the theoretical background underlying our current research. This consists in the development of behaviour-based knowledge systems, for closing the gaps between behaviour-based and knowledge-based systems, and also between the understandings of the phenomena they model. We expose the requirements and stages for developing behaviour-based knowledge systems and discuss their limits. We believe that these are necessary conditions for the development of higher order cognitive capacities, in artificial and natural cognitive systems.

cs.AI↗

arXiv2002

Complex Philosophy

We present several philosophical ideas emerging from the studies of complex systems. We make a brief introduction to the basic concepts of complex systems, for then defining "abstraction levels". These are useful for representing regularities in nature. We define absolute being (observer independent, infinite) and relative being (observer dependent, finite), and notice the differences between them. We draw issues on relative causality and absolute causality among abstraction levels. We also make reflections on determinism. We reject the search for any absolute truth (because of their infinity), and promote the idea that all comprehensible truths are relative, since they were created in finite contexts. This leads us to suggest to search the less-incompleteness of ideas and contexts instead of their truths.

nlin.AO↗

arXiv2002

A Study on the Relevance of Information in Discriminative and Non-Discriminative Media

In this paper we compare the relevance of information obtained from "discriminative" media and from "non-discriminative" media. Discriminative media are the ones which accumulate and deliver information using a heuristic selection of it. This can be made by humans, or by artificial intelligent systems, exhibiting some form of "knowledge". Non-discriminative media just collect and return information without any distinction. This can also be made by humans or by artificial systems, but there is no "knowledge" involved in the process. We ranked the words occurring in an edited electronic publication specialized in complex systems research, and we found that they approximate a modified Zipf distribution. We compared occurrences of representative words from the distribution with the occurrences in non-discriminative media. We found that a non-discriminative medium (Google) has a higher variance from of our original distribution than a semi-discriminative one (NEC Research Index), even when both appear to have their own modified Zipf distribution. We conclude that discriminative media have a higher efficiency rating, at least in the area in which they specialize, than non-discriminative media. Using the same search method, the discriminative media should deliver more relevant information. This relevancy also depends on the skills of the user, but non-discriminative media are more sensitive to poor searching skills, as there is a higher probability of delivering irrelevant information. This leads us to suggest the incorporation of intelligent classifications in different media (such as the ones suggested by the Semantic Web project), in order to increase the relevance of the delivered information.

nlin.AO↗