Towards Detecting Group Identities in Complex Artificial Societies

Togelius

Proceedings of the 15th Annual Conference Companion on Genetic and Evolutionary Computation

2013

Self Cite

We present a computational framework capable of inferring the existence of group identities, built upon social networks of reciprocal friendship, in Complex Adaptive Artificial Societies (CAAS) by solely observing the flow of interactions occurring among the agents. Our modelling framework infers the group identities by following two steps: first, it aims to learn the ongoing levels of cooperation among the agents and, second, it applies evolutionary computation, based on the learned cooperation values, to partition the agents into groups and assign group identities to the agents.Experimental investigations, based on CAAS of agents who interact with each other by means of the Ultimatum (or Bargain) Social Dilemma Game, show that a cooperation learning phase, based on Reinforcement Learning, can provide highly promising results for minimising the mismatch between the existing and the inferred group identities. The proposed method appears to be robust independently of the size and the ongoing social dynamics of the societies.

Section: Discussion and Future Workmentioning

confidence: 99%

Section: Cooperation Learning Via Reinforcement Learningmentioning

confidence: 99%

Section: Cooperation Learning Via Reinforcement Learningmentioning

confidence: 99%

“…whilst pHAP (ai, aj) is obtained as the linear combination of pAP and the normalised entropy [11] of the offers:…”

Section: Cooperation Learning Via Reinforcement Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Interaction-based group identity detection via reinforcement learning and artificial evolution

Togelius

Proceedings of the 15th Annual Conference Companion on Genetic and Evolutionary Computation

2013

Self Cite

Validating Generic Metrics of Fairness in Game-Based Resource Allocation Scenarios with Crowdsourced Annotations

Martínez

Transactions on Computational Intelligence XIII

2014

Self Cite

Being able to effectively measure the notion of fairness is of vital importance as it can provide insight into the formation and evolution of complex patterns and phenomena, such as social preferences, collaboration, group structures and social conflicts. This paper presents a comparative study for quantitatively modelling the notion of fairness in one-to-many resource allocation scenarios-i.e. one provider agent has to allocate resources to multiple receiver agents. For this purpose, we investigate the efficacy of six metrics and cross-validate them on crowdsourced human ranks of fairness annotated through a computer game implementation of the one-to-many resource allocation scenario. Four of the fairness metrics examined are well-established metrics of data dispersion, namely standard deviation, normalised entropy, the Gini coefficient and the fairness index. The fifth metric, proposed by the authors, is an ad-hoc context-based measure which is based on key aspects of distribution strategies. The sixth metric, finally, is machine learned via ranking support vector machines (SVMs) on the crowdsourced human perceptions of fairness. Results suggest that all ad-hoc designed metrics correlate well with the human notion of fairness, and the context-based metrics we propose appear to have a predictability advantage over the other ad-hoc metrics. On the other hand, the normalised entropy and fairness index metrics appear to be the most expressive and generic for measuring fairness for the scenario adopted in this study and beyond. The SVM model can automatically model fairness more accurately than any ad-hoc metric examined (with an accuracy of 81.86%) but it is limited by its expressivity and generalisability.

Artificial evolution for the detection of group identities in complex artificial societies

Togelius

2013 IEEE Symposium on Artificial Life (ALife)

2013

Self Cite

This paper aims at detecting the presence of group structures in complex artificial societies by solely observing and analysing the interactions occurring among the artificial agents. Our approach combines: (1) an unsupervised method for clustering interactions into two possible classes, namely ingroup and out-group, (2) reinforcement learning for deriving the existing levels of collaboration within the society, and (3) an evolutionary algorithm for the detection of group structures and the assignment of group identities to the agents. Under a case study of static societies-i.e. the agents do not evolve their social preferences-where agents interact with each other by means of the Ultimatum Game, our approach proves to be successful for small-sized social networks independently on the underlying social structure of the society; promising results are also registered for mid-size societies.