Categoría: Congresos

Publicado el

[CostAT] Coherence-based argumentation models for normative agents

by Sindhu Joseph

Abstract :
In this talk coherence-based models are proposed as an alternative to argumentation models for the reasoning of normative agents and normative deliberation. The model is based on Thagard’s theory of cognitive coherence and exploits the coherence relations that exist between claims and conclusion of arguments. A coherence-based model is intended to introduce more flexibility in the process of deliberation and agreement generation among normative agents. The basic coherence philosophy and what makes it interesting in the context of normative agents that deliberate to regulate a domain of interest are discussed.

This paper shows the application of coherence models to an argumentation model in a normative, regulated environment. I’m interested not in tris particular application, but in the coherence theory (Thagard).

Coherence estudies associations between pieces of information. It tríes to separate information in sets that mutually support the data. In some way, it can be consideres as a constraint satisfacción problem.

Different types of coherence can be identified: deductive, explanatory, deliberative, analoogous or conceptual, depending on the type of information. The Thagard model is a model of deductive coherence. It can be considered as a constraint satisfacción problem. But the main difference is that it does not try to maximize the partition (not the optimal -it is not needed to find a solution-)

Coherence applied to argumentation sees positive relates info as supporting arguments and negative weights as attacks to a claim.

Problem (general) How the coherence weights are calculated? Well, it is addressed in the questions: depends (roughly) on the number of arguments supporting a hypothesis.

Something interesting in the conclusiones: it can model different tupes of agente (utility maximizares, norm abiders, altruistic…) What about diferentt personalities? And a possibility for us: introduction of contexto as part of the future work.

More infomration, read Sindhu Jospeh PhD. thesis, «Copherence-Based Computational Agency«

Publicado el

Consensus Networks as Agreement Mechanism for Autonomous Agents in Water Markets

Es el título de nuestro paper en las Jornadas que organiza el \(im^2\) (Instituto de Matemática Multidisciplinar) de la UPV: Mathematical Models for Addictive Behaviour, Medicine & Engineering. El tema es el uso de redes de consenso para alcanzar acuerdos de forma descentralizada, aplicado en concreto a problemas de gestión de recursos hídricos. A continuación te dejo el resumen (en inglés) y las trasparencias de la presentación. En cuanto esté publicado dejaré también la referencia completa al artículo y, si puedo por temas de licencia, el enlace.

Abstract

The aim of this paper is to present a way of share opinions in a decentralized way by a set of agents that try to achieve an agreement by means of a Consensus Network, allowing them to know beforehand if there is possibilities to achieve such an agreement or not.
The theoretical framework for solving consensus problems in dynamic networks of agents was formally introduced by Olfati-Saber and Murray (2004). The interaction topology of the agents is represented using directed graphs and a consensus means to reach an agreement regarding a certain quantity of interest that depends on the state of all agents in the network. This value represents the variable of interest in our problem.

A consensus network is a dynamic system that evolves in time. Consensus of complete network is reached if and only if \(x_i = x_j \forall i, j\). Has  been de demonstrated that a convergent and distributed consensus algorithm in discrete-time can be written as follows:

\(x_i(k+1)=x_i(k) + \varepsilon \sum_{j \in N_i} a_{ij}(x_j(k)-x_i(k))\)

where \(N_i\) denotes the set formed by all nodes connected to the node i (neighbors of i). The collective dynamics of the network for this algorithm can be written as \(x(k+1)=Px(k)\), where \(P=I-\varepsilon L\) is the Perron matrix of a graph with parameter \(\varepsilon\). The algorithm converges to the average (or other functions) of the initial values of the state of each agent and allows computing the average for very large networks via local communication with their neighbors on a graph.
The convergence of this method depends on the topology of the network and its convergence is usually exponential. But sometimes it not needed to reach a final agreement on a concrete value. This proposal uses consensus networks to determine if an agreement is possible among a set of entities. Agents can leave the agreement if its parameters are out of the expected bounds, so the consensus network can be used to detect the candidate agents to be members of the final agreement. All this process is solved in a self-organized way and each individual agent decides to belong or not to the final solution.
To show the validity of the present approach, a water market is presented as case of study. The water market is a case of complex social-ecological system (SES), where centralized and hierarchical approaches trend to fail and self-organized solutions seems to be more sustainable in the long term (Ostrom, 2009). In general, agreements related to natural resource management involve very complex negotiations among agents. Water demands and regulation is a very complex distributed domain appropriated for MAS.
An important question is if this kind of markets requires some regulation or not. From an exclusively economic point of view the dominant strategy for agents in deregulated markets is not cooperative because each agent wants to maximize exclusively his payoff, and therefore they are not interested in the global and socially efficiency of the natural resources.

Publicado el

[AT workshop] Session 4

Reputation and confidence for artificial intelligent entities. A cognitive approach
(Jordi Sabater)

Trust deals with uncertainly and risky situations. A little difference: reputation (very similar) is one of hte mechanism to build trust and it is a social element. How it is used in a computer-based systems? Three layers (approaches): security, institutional and social. Trust and reputation are meaningful in the social approach. If we have a storngly ruled system (institutional approach) we do’t need trust, just to follow the rules. Then, a cognitive model of reputation is needed.

A social evaluation is the evaluation by a social entity of some property (mental, physical or social) related with been social. Reputation is then a voice (something that is said) about a social property. But agents do not have to beleave this reputation measures: agents (as people) has no responsibility about spreading social evaluations. When people believes what other people sais, then reputation matches with image (what an agent believes in, consideres as true facts).Reputation means communication and gossiping is the channel used to transmit reputation measures. Images and reputation are based on facts, which have two measures: value and strength -> repage mechanism.

This repage cognitive computational model has to be inserted in an agent. It is important that (i) reputation model can be isolated from other reasoning mechanisms (planners, decision making tools); and (ii) be proactive: do not wait to be asked about reputation, but provide information to the rest of elements. Using a BDI (beliefs, desires and intentions) model with multicotext logics and bridge rules to integrate the context of teh repage mechanism into the context of beliefs, desires and intentions. In the logic, the difference between images and reputation is a ¿reified? difference. An argumentation model is used

Psychopharmacology of agreement
(Adolf Tobeña)

There’s lots of corrdination, obbidion, … but few agreements among humans. ANd the second point of the speech is that humans need drugs. And these facts «llevan» to psychiatric aspects of agreement: why patients are more trending to cooperate/agree after been treated?

Usually, xanthines (caffeine, tobacco) are present during negotiations and bargaining processes. 5 years ago was demostrated that oxitocin increases trust in humas. Furthermore, they observed that participants trend to not change the trusting behaviour even after knowning they had been betrayed (50% trials) and the brain was actually don’t responding as been betrayed (e.g. activity in brain areas related with dissgust).

booster drugs for agreements (prosocial, protrust)

  • alcohol, cannabinoids
  • xanthines, nicotine
  • oxytocine, prolactine, NPY
  • estrogens

and antiagreement drugs are (indice paranidogenic, autistic and antisocial behaviors)

  • cocaine, amphetamines
  • LSD, mescaline, psilocibine
  • androgens

But they’ve observe that testosterone had a possitive effect on human bargaining behavior…. and they did it on women!!!! They shown that one sunlingual dose os testosterone in women cause a substantial increasein fair bargaining, reducing cinflics and increasing efficiency on social interactions. ANd usinga placebo they demonstrate that was a real effect (the believed testosterone group behaves as the group without testosterone. And in men? Other group showed that high levels of testosterone (natural measuring) reject low (unfair) ultimatum game offers: $5/$40. Testosterone has influence in how the rest of the people consider others as leaders. Testosterone redcuces conciuos detrection of signals (face expressions) serving social correlations ->  a high probability of entering into a fight is related with risk/venturesome behavior (you accept more faces as neutral)

Blogged with the Flock Browser
Funciona gracias a WordPress