Jordi Delgado

Extracting reputation in multi agent systems by means of social network topology

The problem of calculating a degree of reputation for agents acting as assistants to the members of an electronic community is discussed and a solution presented. Usual reputation mechanisms rely on feedback after interaction between agents. An alternative way to establish reputation is related with the position of each member of a community within the corresponding social network. We propose a method based on this idea, which is also used by well-known ranking algorithms, discuss its properties as well as experimental results and compare them to other reputation mechanisms for electronic communities supported by agents. The method proposed uses only local information in order to extract reputation and it is able to adapt automatically to the topology of the network or graph.

Emergence of social conventions in complex networks

The emergence of social conventions in multi-agent systems has been analyzed mainly in settings where every agent may interact either with every other agent or with nearest neighbours, according to some regular underlying topology. In this note we argue that these topologies are too simple if we take into account recent discoveries on real networks. These networks, one of the main examples being the Internet, are what is called complex, that is, either graphs with the small-world property or scale-free graphs. In this note we study the efficiency of the emergence of social conventions in complex networks, that is, how fast conventions are reached. Our main result is that complex graphs make the system much more efficient than regular graphs with the same average number of links per node. Furthermore, we find out that scale-free graphs make the system as efficient as fully connected graphs.

Pattern formation and optimization in army ant raids

Army ant colonies display complex foraging raid patterns involving thousands of individuals communicating through chemical trails. In this article we explore, by means of a simple search algorithm, the properties of these trails in order to test the hypothesis that their structure reflects an optimized mechanism for exploring and exploiting food resources. The raid patterns of three army ant species, Eciton hamatum, Eciton burchelli, and Eciton rapax, are analyzed. The respective diets of these species involve large but rare, small but common, and a combination of large but rare and small but common food sources. Using a model proposed by Deneubourg et al. [4], we simulate the formation of raid patterns in response to different food distributions. Our results indicate that the empirically observed raid patterns maximize return on investment, that is, the amount of food brought back to the nest per unit of energy expended, for each of the diets. Moreover, the values of the parameters that characterize the three optimal pattern-generating mechanisms are strikingly similar. Therefore the same behavioral rules at the individual level can produce optimal colony-level patterns. The evolutionary implications of these findings are discussed.

Small-world behaviour in a system of mobile elements

We analyze the propagation of activity in a system of mobile automata. A number ρLd of elements move as random walkers on a lattice of dimension d, while with a small probability p they can jump to any empty site in the system. We show that this system behaves as a Dynamic Small World (DSW) and present analytic and numerical results for several quantities. Our analysis shows that the persistence time T* (equivalent to the persistence size L* of small-world networks) scales as T* ~ (ρp)−τ, with τ = 1/(d + 1).

Universal computation in fluid neural networks

Fluid neural networks can be used as a theoretical framework for a wide range of complex systems as social insects. In this article we show that collective logical gates can be built in such a way that complex computation can be possible by means of the interplay between local interactions and the collective creation of a global field. This is exemplified by a NOR gate. Some general implications for ant societies are outlined.

A Ranking Algorithm Based on Graph Topology to Generate Reputation or Relevance

We introduce NodeRanking, a new mechanism for ranking the importance of nodes in a graph. Ranking nodes in a graph is not a new problem and some solutions do exist in the literature, such as the well-known algorithms Pagerank, HITS, and SALSA. Our main contribution consists in the fact that the introduced algorithm uses only local information. Thus, the algorithm is truly distributed and it does not need any knowledge of the whole graph. Furthermore, NodeRanking adapts itself to the graph topology, hence, no setting-up process is required. The quality and adaptability of NodeRanking is tested on different graphs (a real social network and a scale-free graph) with different topological properties Finally, we show how the algorithm may be applied either to extract the relevance of Web pages or to infer the reputation of members of a community.

Noise induced transitions in fluid neural networks

Abstract We study the autosynchronization of temporal patterns of activity in Leptothorax ants, using to model this phenomenon the formalism of fluid neural networks. It is known that autosynchronization is involved in complex phenomena observed in ant colonies, such as task allocation and mutual exclusion. We have numerical evidence that this happens at the critical point of a noise induced transition.

Characterizing turbulence in globally coupled maps with stochastic finite automata

Abstract We describe a method to discriminate between ordered and turbulent behavior in a general class of collective systems known as Globally Coupled Maps (GCM). Our method is able to discover an unknown small ordered region inside the turbulent phase of GCM parameter space. The computational nature of the method is the main novelty of our approach; it is another example of how measures based on computational notions of structure may provide new information in the study of dynamical systems.

Weighted contrastive divergence

Learning algorithms for energy based Boltzmann architectures that rely on gradient descent are in general computationally prohibitive, typically due to the exponential number of terms involved in computing the partition function. In this way one has to resort to approximation schemes for the evaluation of the gradient. This is the case of Restricted Boltzmann Machines (RBM) and its learning algorithm Contrastive Divergence (CD). It is well-known that CD has a number of shortcomings, and its approximation to the gradient has several drawbacks. Overcoming these defects has been the basis of much research and new algorithms have been devised, such as persistent CD. In this manuscript we propose a new algorithm that we call Weighted CD (WCD), built from small modifications of the negative phase in standard CD. However small these modifications may be, experimental work reported in this paper suggests that WCD provides a significant improvement over standard CD and persistent CD at a small additional computational cost.

Whitehead Problems for Words in

We solve the Whitehead problem for automorphisms, monomorphisms and endomorphisms in Zm Fn after giving an explicit description of each of these families of transformations.