Sistema para la evaluación de la confianza en redes distribuidas

El pasado lunes se presentó el artículo ya anunciado en el anterior post. Como conclusiones, discutimos sobre:

  • La necesidad de confianza entre participantes dentro de una red distribuida, principalmente enfocada a actividades P2P.
  • La manera de medir esa confianza, de modelarla y los beneficios de hacerlo.
  • El obtener las medidas de confianza bien por observación o bien por recomendaciones, y las condiciones que se han de cumplir para una correcta propagación de los valores.
  • Conocimiento de los distintos ataques de: bad mouthing, on-off, Sybli, new user creación de conflictos.
  • Discusión sobre los módulos que compondrían un sistema que gestiona la confianza.
  • Comentarios sobre las gráficas de resultados y los posibles beneficios de usar este sistema cuando se presentan los ataques mencionados.

La presentación puede consultarse a continuación:

Advertisements

Paper Seminar: A Trust Evaluation Framework in Distributed Networks: Vulnerability Analysis and Defense Against Attacks

El lunes que viene, 21 de enero, se discutirá sobre el trabajo A Trust Evaluation Framework in Distributed Networks: Vulnerability Analysis and Defense Against Attacks, colaboración de los departamentos de Ingeniería Electrónica y de Computación de las universidades de Rhode Island y Maryland. En él se propone una manera de aumentar la seguridad en arquitecturas de red distribuidas, basándose en una medida cuantitativa de la confianza en cada entidad participante. Será en la Sala de Reuniones de la ETSIIT, estáis todos invitados.

Evolutionary Algorithms in Heterogeneous Nodes

Today I presented a brief talk about some papers about the usage of heterogeneous computers for distributed EAs:

All these ideas (and new ones) are being applied in our Service Oriented Architecture for Evolutionary Algorithms, we hope to show interesting results soon!

Here is the presentation (in Spanish).

Quién le dice qué a quién

Para comenzar de nuevo con la serie de paper seminars mañana día 23 de septiembre hablaremos de uno de los pocos trabajos que estudia la dinámica de Twitter, que está escrito nada menos que por la mitad de Watts-Strogatz: Who says what to whom on Twitter, donde nos descubre si Oprah tiene más influencia que Lady Gaga o es al revés y nos enseña una metodología curiosa para clasificar a los usuarios de Twitter en diferentes tipos.
Como es habitual, el seminario será a las 12:30 de la mañana en la sala de reuniones de la ETSIIT

Dynamic Control in Evolutionary Algorithms

Last Friday, in our weekly meeting, the paper by Di Tollo et al. “From Adaptive to More Dynamic Control in Evolutionary Algorithms” was presented and discussed. This work in centered on the adaptation of application rates of different types of crossover. A performance function is defined that takes into account the quality of solutions and diversity generated by each crossover. By varying a user-defined variable (teta), the importance of each factor can be regulated in order to set the desired compromise between quality and diversity (which gives rise to the idea of applying a multi-objective approach here). Then, after credit assignment (for each crossover), an operator is selected by Probability Matching (PM) or Multi-Armed Bandit (MAB) strategies.

For testing the proposed scheme, the authors define a framework with 20 different crossover operators of which the main characteristics are known (i.e., whether they favor intensity/exploitation or diversification/exploration). The system is applied to SAT problems. Several conclusions are drawn from those simple experiments. First, the type of SAT problem greatly influences the behavior of the system, as well as the criteria used to compute the performance of the operator and the selection strategy (PM or MAB). That is, setting teta to a hypothetical compromise value between intensification and diversification leads to a variety of different behaviors and not necessarily to that expected compromise.

The second part of the experiments is focused on the dynamic variation of teta. The authors conclude that the variation strategy influences the behavior of the algorithm and the progress of the system. They also conclude that, in fact, it is possible to favor diversity-oriented crossover or quality-oriented crossover by tuning teta. That is, it is possible to control the desired features by changing the teta value during the search.

Paremeter Tuning (and a bit of Control) in EAs

As I promised, I gave a talk about Parameter Setting (Control and Tuning) to other GeNeura members (and invited!). This talk is based mainly in a cool (and easy to read!) work of A. E. Eiben and S.K. Smit (download draft here), but recently a new version with more information and stuff has been published in Journal of Swarm and Evolutionary Computation (you can download the paper directly from Eiben’s webpage here). The first one is more introductory but the second delves into the more technical aspects.

The slides are here (sorry, they are in Spanish :(  )

Also you can download the Keynote (.key) file from here. It includes EPIC mega-hella-awesome-mac-transitions! (yeah, I have to admit I was willing to give a talk just to show my new Mac’s awesomeness, sorry).

Pablo

Talk about parameter setting in Evolutionary Algorithms

Next friday I’ll give a talk about parameter control and parameter tunning (focusing in the latter) in EAs. I’ve read several papers of professor A. E. Eiben, about this interesting area and I will explain the differences among several kinds of parameters, different classification of settings and the state of the art with examples. You are all invited.

Date: Friday 15th. 12:00 pm.

Place: Sala de Juntas, ETS. Ingenierías en Informática y Telecomunicación. Universidad de Granada.

Pablo.