Day 1 – Monday, October 5, 2009
International Conference on Computational Collective Intelligence – Semantic Web, Social Networks & Multiagent Systems
For official information check website: ICCCI 2009. Conference statistics: 212 submissions, 71 accepted, 21 accepted to be published in LNAI, 25 countries, 3 days, 3 concurrent tracks. Next year ICCCI2010 will be September in Taiwan.
Introduction and opening
Speakers: Ngoc Thanh Nguyen (Wroclaw), Ryszard Kowalczyk (Melbourne), Swiatek, Zbigniew Huzar (Wroclaw)
First plenary session: Roman Slowinski, Poznan University of Technology, PAN
Introduction to Rough Sets, see Zdzislaw Pawlak (1926-2006)
The problem of inconsistencies in the data sets. The proposed approach is to use rough sets.Important concepts: granules, indiscernible objects, lower approximations, higher approximations, Tools: Rose, 4eMka, JAMM, jMAF, see http://idss.cs.put.poznan.pl/site/software.html
First session, stream B, Hedonic Coalition Games
Helena Harlotta Kainanen: Coalition formation, metropolitan algorithm, stable coalition structures. Comparison of AMS and RII algorithms. Stochastic Local Search proposed as future work.
Luca Longo, Stephen Barrett, Pyerpaolo Dondio: Foraging theory for modelling user behaviour when browsing the web three generation of web search: content, link analysis, behaviour analysis (collaborative ranking). Using the behaviour of users to estimate the usefulness of web pages. User-behaviour is observed and measured via proprietary web-browser plugin (Firefox), which records all the actions the user undertakes. All actions are weighted in an arbitrary way into a single effort measure, which is then used to estimate/predict future user behaviour.
Krissada Maleewong, Chutiporn Anutaryya, Vilas Wuwonges: Semantic Argumentation based Model for Collaborative Knovwlege Creation and Sharing System. Study of question-answer systems.
Szymon Lukasik, Slawomir Zak: Firefly Algorithm for Continuous Constraned Optimization Tasks
Note: see the book: Nature-inspired Metaheuristic algorithms by Xin-She Yang (Luniver Press, 2008)
Marcin Gorawski, Ewa Pluciennik-Psota: Distributed Data Mining Methodogy with Classification Model Example
Plenary session: Piotr Jedrzejowicz, A-Teams
A-Teams and their applications, Gdynia Maritime University
- Talukdar, S. et al Journal of Heuristics 4 295-321 (1998). Talukdar invented the concept of A-Teams.
- Akkiraju, R. et al, Applied Intelligence, 2001
- Meneses et al 2008 (next generation A-Teams framework)
- Jedrzejowicz, Werzbowska, 2006 (JABAT, JADE-based A-Teams environment.
- Barbucha, 2006, 2008
- Czarnowski 2009
Start by generating the example solutions plus shared memobry. writing operations. reading operations. Agents are asynchronous, heterogenous, and organised into star-like topology. There is an elected supervisor (controller) chosen per assembly. This central agent works as facilitator. Disadvantages: facilitator is a single-point-of-failure, communication bottleneck, communication overheads. Related architectures:
- Architectures for evolutionary and coevolutionary multi-agent systems (EMAS, CoEMAS).
- Adaptive memory programming (AMP).
- Decision Making Library (Agent-DMSL), written in Java
- Swarm intelligence systems
First generation tools: IBM A-Team written in C++ (1999), Matlab-based A-Teams shell (Talukdar). Constructors, improvers, destroyers. A-Teams theoretical analysis vey rudimentary. Third-generation should allow self-optimisation mechanisms.
Second session, stream A
Zulfiqar A. Memon, Jan Treur: Designing Social Agents with Empathic Understanding, Vrije University of Amsterdam. Theory of Mind (mindreading). Two perspectives: simulation-theory and theory-theory. Empathy: the ability to put oneself into the mental shoes of another person to understand his/her emotions and feelings (Goldman, 1993). 3 mental states of empathy for the observing person: feeling without understanding, understanding without feeling, both, feeling and understanding. Modelling empathy in MAS, by replicating internal processing and in-mind and body-based mental loops (see the book below).
Read the book: Antonio Damasio (1999)
Jaroslaw Was, AGH Krakow: MAS in Pedestrian Dynamics Modeling. Review of the pedestrian models: Dijkstra 1990, SBBPedis, Situated CAs, Navigation dynamics, Alpsim.
Grzegorz Skorupa, Radoslaw Katarzyniak: Towards a Model for Extraction of Possible Worlds and Accessibility. Relation from Cognitive Agent. Discussion about possible worlds. No implementation yet, the authors are looking for sufficiently powerful algorithms that could be used to analyze the search space.