Psico-a.eu

psico-a.eu

Ted Szuba: Chaos and collective intelligence

Tadeusz (Ted) Szuba is a pioneer of the formal study of the concept of collective intelligence (CI). Collective intelligence is group intelligence that emerges from the competition or collaboration of many individuals. It may involve Biology, Sociology, Computation or Business. In 2001, Dr. Skuba proposed a mathematical model for the phenomenon assuming to be a random, distributed and unconscious process. Recently, Professor Skuba has studied Chaos as one of the basic properties for CI. For instance, in bacterial colonies emerges CI based on mutual observation of metabolism products and the exchange of DNA. This phenomenon may vanish if a form of Chaos is not maintained. Professor Szuba asks himself what should the measure and form of Chaos be in the social structure. In case of society, Szuba speaks in terms of parallel Chaos. When individual intelligence of a being is poor, e.g., ants, bees, etc., it must be massive parallelism. It is obvious that increasing the organization of a social structure reduces chaotic behavior of individuals but reduces ability to solve unexpected problems, which a social structure can confront. In fact, panic behavior has a chaotic nature. In case of bacteria, for instance, this is some kind of mutation, while in case of human beings it can take any form of escape. According to Szuba, societies frequently add a chaotic element in an artificial way. For example, the concept of lottery giving an illusory chance to people, to change their life. Szuba concludes that the ability of the societies to solve specific class of problems, expressed through CI becomes non-Chaotic, despite a lot of individual Chaos inside the society.

Cynthia Breazeal and the Personal Robots Group

Cynthia Breazeal directs the Personal Robots Group at the MIT, being a pioneer of social robotics. She develops socially intelligent robots interacting and communicating with people. She wishes to help people of all ages contributing to their quality of life, in aspects such as health or learning and education. 

Socially aware robots require a broad range of capabilities in order to work together with human beings. These robots must be able to recognize intent and action supposing a challenge to the robotics community. It is a very difficult task to examine highly complex environments in the number of states, transitions and interactions. Machine learning methods like Reinforcement Learning or POMDPs do not scale to high-dimensional state-action spaces. Breazeal and her group think to endow the robots with human-like behaviour through generating functionally capable data-driven behaviours sourced from human behaviour. For it, they use of plan networks and case-based planning simulating crowdsourcing behaviour. So it is harnessed the wisdom of the crowd. They evaluate the resulting autonomous robot behaviours using a real-world reproduction of an online game environment and obtaining interesting results which optimize the results  about social interaction between humans and robots, obtained by traditional methods

My-Pet-Our-Pet

My-Pet-Our-Pet (see the article by Chen et al., 2007-http://www.petpartners.org/document.doc?id=289-) is an animal companion system based on open learner models as animal companions. The learners are grouped into several teams, in which each learner is surrounded by two kinds of animal companions, an individual animal or My-Pet, and a team animal companion or Our-Pet. My-Pet encourages users to learn individually online subjects through nurturing the pet character. The repertoire of activities in the system consists of four modes: pet nurturing, individual learning, game competition and group discussion. In an experiment in a classroom of one elementary school in Taiwan, 31 students were motivated and engaged in the process of raising their pets, and most of them (29 students) increased effort to improve their learning. Open learner models as animal companions benefit student´s learning.

Social Cognition and Collective Intelligence

Collective intelligence is the capacity of human collectives to engage in intellectual cooperation. There exists an interesting correlation between collective intelligence and the degree of the human development. Some months ago it was held a dialogue in MIT TV between Andrew Lo (MIT), Martin Nowak (Harvard), Alexander Pentland (MIT) and Rebecca Saxe (MIT). Thomas Malone moderated the dialogue. The interested reader may find the link here: http://video.mit.edu/watch/social-cognition-and-collective-intelligence-7792/

Living Technology and Artificial Intelligence

Life has always been a source of inspiration for technology. Nowadays protocells are being created including artificial plasmids that produce functional proteins. In fact, electronic chemical cells present information processing between the chemistry and electronics. Digital organisms combine self-assembly with genetic descriptions providing a basis for using evolvable technology.

What living technologies will be like in the future? We can imagine organisms with synthetic genomes to produce hydrogen in fuel cells. Also we can imagine self-assembling, self-correcting, and adaptive software solving many problems of lack of robustness in software. Autonomous and adaptive robots  will enable the elderly to live independently and with greater security.

The emergence of living technology will create new social processes. We can expect purely artificial technology to acquire life´s properties and outperform the current technology. This process will be a singular event in human history.

Pedagogical agents and Artificial Intelligence

Pedagogical agents, that is, animated, life-like characters in a computer-based learning environment are attracting research in Education. Students learn lessons while interacting with agents designed to apply instructional protocols. In fact, learners can receive from the agent verbal messages tipically provided by humans. In a certain sense, a learner might be able to hold a connection with that agent, just as in a classroom. It is recommended that a pedagogical agent look human-like looking for a peer-to-peer interaction over teacher-to-learner interaction. The value of interaction with equally able peers generates the cooperation of equal partners fostering thinking, intellectual development and even affect. By this, researchers have attempted to simulate this interaction in tutoring systems. For instance, Biswas and colaborators at the Vanderbild University with the project Betty´s brain. Betty is a kind of anthropomorphized learning companion building social relations with the student. A pedagogical agent is not a conventional computer-based approach. Traditional human-computer interaction approaches haven´t studied how to support social interaction. Antropomorphized learning companions are just the possibility for such that exploration. It is obvious that the social interaction with the teachers and the learner´s perception of them play an important role in influencing the motivation and the learner´s efficacy. So the social interaction with learning companions in computer-based environments.

PARO or the therapeutical use of robots

Eight years ago, PARO, a therapeutical robot developed by Takanori Shibata, began to be sold commercially. PARO has the appearance of a seal but also presenting characteristics of human babies. It has been used with dementia patients in Japan, Italy or Germany. In an experimental study, half of the dementia patients that interacted with PARO improved their brain activity. PARO can successfully create affective bonds with people. PARO is the best proof that for humans is easier to interact with robots different from what we experience. Humans feel uncomfortably interacting with mechanical objects very similar or that cannot be distinguished from themselves. Psychology uses toys in therapy and we can create artificial seals and project upon them the feelings that we ascribe to real seals. But being ascribed a property and having it are not the same thing. The uncanny hypothesis holds that when human replicas look and act almost, but not perfectly, like actual human beings, it causes revulsion among human observers. This is an interesting lesson for the development in robotics of the imitation myth because people interact better with a robot that does not provoke expectations of how it should behave.



Duncan Luce (1925-2012): in memoriam

(Powerpoint by Jean-Claude Falmagne in the University of Navarra,

 remembering to Duncan Luce (August, 30, 2012). (Photo by Carlos Pelta)

Duncan Luce died August, 11, 2012, in Irvine (USA). Professor Luce has been one of the most renowned mathematical psychologists in 20th century. Pioneer of the quantitative formalization in Psychology and of the rigorous analysis of the Theory of Measurement, I remember the deep impact of his book Individual choice behavior on my youhtful mind. It was fascinating to see how Luce´s version of the principle of the independence of irrelevant alternatives was exploited in new and surprising directions. Choice probabilities for different sets of alternatives satisfying the principle of independence of irrelevant alternatives might  be consistently defined independently of the actual choice set.

I was lucky to shake hands with Professor Luce in the the "2011 Meeting of the European Mathematical Psychology Group" at  Paris one year ago. Requiescat in pace, Professor Luce.

EMPG 2012: Mathematical Psychology in the University of Navarra

                      (Access the University of Navarra) (Photo by Carlos Pelta)                                                

The annual Meeting of the "European Mathematical Psychology Group" (EMPG 2012) has been held at the University of Navarra (Pamplona, SPAIN), 29-31 August, 2012, and has been a great success. The key of this success has been Professor Christine Choirat, chair of the Meeting, whose excellent organization has been acknowledged by all concerned. Thanks to Professor Choirat and to the colaborators of the Economics Faculty and thanks to the University of Navarra for receiving so favourably this Meeting. The next Meeting will be celebrated in Postdam (Germany).

Between some of the most interesting talks I will emphasize the following:
On Wednesday, Professor Jacqueline J. Meulman  (Leiden University) spoke in a plenary talk about the new approach developed in Leiden called nonlinear multidimensional data analysis. Professor Noventa (University of Padua) presented in colaboration with Professors Stefanutti and Vidotto, an analysis of item response theory and Rasch models based on the most probable distribution method. Professor Budescu (Fordham University) analyzed test-taking behavior showing that penalties for incorrect answers have detrimental effects for both Test-Takers and Test-Makers. Professor Hudry (Telecom ParisTech) showed the NP-hard nature of the computation of a linear order or of a complete preorder under remoteness conditions. Professor Doignon (Brussels) described a geometric interpretation of therelationship between the probability distribution on knowledge states and the derived distribution on response patterns. Professor Núñez-Antón in colaboration with Professors Arostegui and Quintana, talked about the comparation of the outcomes of eight techniques of logistic regression for studying data from depressive patients. Professor García-Pérez (Complutense University of Madrid) proposed a solution for the problem of residual analysis in contingency tables. Professor Carlos Pelta presented his computational system based on agents (PSICO-A) for teaching Psychology and the journey was closed by Professor Farina (Siena) with an experiment about choice reversal in anticipatory feelings.
On Thursday, Rutherford (Keele University) developed a plenary talk about methodological problems concerning to the hypothesis to be tested in Psychology. Professor Colonius (University of Oldenburg) analyzed his Universal Fechnerian Scaling technique, a method for computing subjective distances among stimuli from their pairwise discrimination probabilities. In the afternoon Professors Erber, Goebel and Nan from Vienna, presented their systems for pattern recognition training by using visual feedback to the amputee.
Undoubtely the most emotive presentation of the Meeting was developed by Professor Jean-Claude Falmagne (University of California, Irvine), pioneer of the Mathematical Psychology and founder member of the European Mathematical Psychology Group. Professor Falmagne remembered the contribution and aspects of the life of the late Professor and friend Duncan Luce, one of the most important mathematical psychologists in 20th century.
Professor Alcalá-Quintana described an extension of her indecision model in psychophysics for producing second choices that are consistent with empirical data obtained under a second-choice paradigm without resorting to the increasing-variance assumption. The model uses a proper two-alternative forced-choice (2AFC) task with a three response format. Just at that moment Professor Laming (University of Cambridge) established a correlation between the 2AFC paradigm and data corresponding to psychophysics.
On the last day Professor Pigozzi (Paris Dauphine) developed in an interesting plenary talk her application of the labeled deductive systems theory to the psychological aspects of the argumentation. Professor Suck (Universuty of Osnabrück) introduced his investigations about set valued random variables and Professor Induráin (UPN) spoke about results concerning to separability properties relative to semiorders (results obtained in colaboration with Professor Estevan and professors Candeal and Gutiérrez-García). Finally Professors Stefanutti, de Chiusole and Spoto (University of Padua) introduces their ideas on the basic local independence model, a restricted latent class model for probabilistic knowledge structures.

European Mathematical Psychology Group Meeting (EMPG 2012)

Organized by Professor Christine Choirat (Universidad de Navarra), and with a Scientific Committee formed by Professors Bouyssou (France), Colonius (Germany), Doignon (Belgium), Falmagne (France and USA), Hudry (France), García-Pérez (Spain), Marchant (Belgium) and Induráin Eraso (Spain), it will be held at the University of Navarra (Pamplona-Spain), the annual Meeting of the European Mathematical Psychology Group (EMPG-2012), from August 29, 2012 until August 31, 2012.

The plenary speakers will be Professors Meulman (Leiden University), Pigozzi (Paris-Dauphine) and Rutherford (Keele University).

The author of this blog, Carlos Pelta, will attend the Meeting with a talk entitled "PSICO-A: A new computational system for learning Psychology". Below the program is available:

https://sites.google.com/site/empg2012/program

Neurosciences Madrid 2012

Organized by Ramón Areces Foundation and coordinated by Professor José Luis Muñiz Gutiérrez (CIEMAT), a series of lectures will present the outlooks from different disciplines and perspectives about Neuroscience from Wednesday, July 4, 2012 until Tuesday, July 5, 2012. Herein I enclose the program of this Meeting:

Program

Coordinated by:
José Luis Muñiz
Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT).
Grupo de Física Médica de Real Sociedad Española de Física (RSEF).

---

Wednesday, 4

09:30 Opening

Raimundo Pérez-Hernández y Torra
Fundación Ramón Areces.

María del Rosario Heras Celemín
Real Sociedad Española de Física. (RSEF).

José Luis MuñizGrupo de Física Médica de la Real Sociedad Española de Física (RSEF).

---

10:00 La neurona de Jennifer Aniston

Rodrigo Quian Quiroga
Department of Engineering. University of Leicester. Reino Unido

---

10:40 Estudio de la conectividad funcional en registros de alta densidad: cuando más es menos

Ernesto Pereda
Departamento de Física Básica. Universidad de La Laguna. Tenerife.

---

11:20 Redes Complejas y epilepsia del lóbulo temporal. Focalizando fuera del foco la causa de las crisis focales

Guillermo Ortega
Hospital Universitario de La Princesa. Madrid.

---

12:00 Break

---

12:30 Sueño, conciencia y complejidad

Enzo Tagliazucchi
Goethe-University Frankfurt. Alemania.

---

13:10 Visualización microscópica del cerebro desde los tiempos de Cajal hasta nuestros días

Javier de Felipe
Centro de Tecnología Biomédica (CTB). Universidad Politécnica de Madrid.

---

13:50 Break

---

16:00 Ritmos cerebrales buenos y malos: estudiando la dinámica neuronal normal y epiléptica

Liset Menéndez de la Prida
Instituto Cajal. CSIC. Madrid.

---

16:40 La física del "dolce far niente", ¿Qué hace el cerebro cuando no hace nada?

Dante Chialvo
Neurophysiology Laboratory. University of California, Los Angeles. EE.UU.

---

17:20 Aplicaciones clínicas de la Imagen Médica en la psiquiatría/salud mental

Celso Arango
CIBERSAM. Hospital General Universitario Gregorio Marañón. Madrid.

---

18:00 Discussion

---

Tuesday, 5

10:00 Neuroimagen por Resonancia en Enfermedades Neurodegenerativas y Neurológicas

Juan Antonio Hernández Tamames
Universidad Politécnica de Madrid.

---

10:40 Neuroimagen estructural en adolescentes con psicosis y autismo

Joost JanssenCIBERSAM. Hospital Universitario Gregorio Marañón. Madrid.

---

11:20 Interfaces Cerebro-Máquina: aplicaciones básicas y clínicas

José M. Carmena
Brain-Machine Interface Systems Laboratory.
University of California, Berkeley. EE.UU.

---

12:00 Break

---

12:30 Estudio de la recuperación del daño cerebral mediante MEG

Nazareth P. Castellanos
Laboratory of Cognitive and Computational Neuroscience. UCM-UPM. Centre for Biomedical Technology.

---

13:30 Inmunología y Sistema Nervioso: conceptos básicos y aspectos clínicos

Juan Antonio García Merino
Hospital Universitario Puerta de Hierro Majadahonda.

---

13:50 Break

---

16:00 Células Madre neuronales. Caracterización electrofisiológica y desarrollo de una terapia celular para el tratamiento del Ictus Isquémico

Josefina María Vegara Meseguer
Universidad Católica San Antonio de Murcia.

---

16:40 Terapia celular en Esclerosis Lateral Amiotrófica: del laboratorio a la clínica

Jonathan Jones*, Mª Carmen Viso, Diego Pastor, Salvador Martínez
(*) Instituto de Neurociencias. Universidad Miguel Hernández. San Juan, Alicante.

---

17:20 Nuevas perspectivas en neurorregeneración: terapia celular aplicada a la discapacidad neurológica

Jesús Vaquero
Servicio de Neurocirugía. Hospital Universitario Puerta de Hierro Majadahonda

PoCoMo: playful social interactions between multiple projected characters

PoCoMo (see http://fluid.media.mit.edu/people/roy/media/PoCoMo-cam-ready-optimized.pdf) consists of two mobile projector-camera systems with computer vision algorithms to support the behaviours of characters projected in the environment. The characters are guided by hand movements and can respond to other characters, simulating a reality of life-like agents. Users hold micro projector-camera devices to project animated characters on the wall and the characters recognize and interact with one another. Extracting visual features from the environment, an algorithm enables operations in a limited resource environment. The system creates games including social scenarios of relating and exchange between to co-located users. The characters are programmed to have component parts with separate articulation and with different sequences based on the proximity of the other characters. The projected characters can respond to the presence and the orientation of one another and acknowledge each other. Also they can trigger gestures of friendship such as shaking hands. Characters can leave presents to one another. Each character has an identity that is represented by the color of its markers. A detection algorithm scans the image by applying a threshold extracting the contour of the figures. The detection algorithm has been implemented in C++ and compiled to a native library. The UI of the application has been implemented in JAVA using ANDROID API. In a future work, the authors (Shilkrot, Hunter and Maes from MIT Media Lab) will integrate markers with the projected content and migrate the application to devices with wider fields of views.

Origin of epistemic structures and computational agents

Adding cognitive structures to the world is a basic adaptive strategy. Chandrasekharan and Stewart (2007) developed an interesting perspective modeling swarms of foraging robots for simulating epistemic structures in agents. Epistemic structures could be structures generated for oneself (for instance, bookmarks), structures generated for oneself and others (pheromones, etc.) and structures generated exclusively for others (warning smells...) (see Chandrasekharan and Stewart, 2007, 331). A key feature of such structures is their task-specifity. According to Chandrasekharan and Stewart (art. cit., 332), organisms sometimes generate random structures in the environment and organisms have a bias to reduce physical or cognitive effort. The authors design a computational simulation consisting of an environment of a 30x30 toroidal gridworld, with one 3x3 square patch representing the agent´s home, and another representing the target. This target can be thought of as a food source. An agen can perform several possible actions: moving randomly; to distinguish between their home and their target (consider models of ant foraging and postulate a home pheromone and a food pheromone) and to modify the environment. The agents have four sensors, two external and two internal and are programmed by a genetic algorithm to evolve foraging behavior in the agents. The fitness function of the genetic algorithm is the inverse of the time measure, which is interpreted as an expression of tiredness. In the simulation, 10 agents foraged at the same time. Initially, the agents behaved randomly. Most agents did not find the target. On average, each agent was completing 0.07 foraging trips every 100 time steps. After a few hundred generations, the agents were completing an average of 1.9 trips in that same period. The result confirmed that the agents were able to systematically make use of their ability to sense and generate structures in the world, on an evolutionary time scale. But is possible for the agents adding structures to the world within their lifetimes? The Q-learning (Watkins, 1989) is the simplest method to perform this task. Using the Q-learning algorithm, 10 agents were ran for 1,000 time steps. To indicate tiredness, it was gave them a reinforcement value of -1 while foraging. When they returned home after finding the target, they were given a reinforcement of 0. By the end of the simulation, agents required only around 150 time steps to make a complete trip (a foraging rate of 0.66 trips in 100 time steps, that is, twice as quick as agents without the structure-forming ability). Even they spent 58% of their time generating structures: epistemic structures generation allowed the agents to complete their foraging task down from 300 time steps. These simulations facilitate an integration of the symbolic and situated views of cognition supporting the extended mind thesis.

Genetic Programming and self-replicating structures in cellular automata

The study of self-replicating structures is a very important field in Artificial Life. Cellular automata have studied two kinds of replicating structures: self-replicating ones and universal constructors. Von Neumann designed complex universal constructors consisting of multiple components. A second type of replicators, self-replicated loops, were studied by Langton, showing that looplike structures used in universal constructors could independently reproduce themselves. The replication process underlying both universal constructors and self-replicating loops, uses a sequential construction in which an arm extends from the parent structure and deposits the child structure. It depends on manually programmed sequential instructions depending on the presence of totalistic transition functions. But today Genetic programming facilitates the evolution of cellular automata given initial structures where cells may have several possible states. Pan and Reggia (2010) obtain replicating structures that are qualitatively different from past manually designed universal constructors and self-replicating loops. As a consequence, it is possible to produce many replicators that vary in just a single property.

To build a Genetic programming system that can program a cellular automata to support self-replication, the authors use trees as data structures ("chromosomes") that represent both structural information and the rules forming the state transition functions. They use a fitness function that generates the ocurrence of multiple copies of an initial structure in the cellular space over time. These self-replicating structures produced by Genetic programming are different from those found in self-replicating loops and universal constructors: there is no an identifiable instruction sequence and no construction arm. An initial structure grows and then divide, making replication very fast. The structures move and it is not possible distinguish between parents and childrens. In some ways, their fissionlike replication process is similar to the splicing during mitosis in biological cells. Replicators can also support the construction of secondary structures as they replicate, either with or without a given initial seed structure. When the replication rules are executed in parallel in each cell, they often employ a strategy that has not been manually created in old constructions. For instance, the moving-wall strategy consisting in a line of replicating structures depositing secondary structures behind it. Thus, Genetic programming is a very powerful tool in the discovery of novel self-replicating structures in cellular automata.

A computational simulation of laws of imitation in Social Psychology

(Spatial prisoner´s dilemma for b=1.6)

I am going to explain the design of a gamed based on the spatial prisoner introducing the three laws of imitation defined by the French Sociologist Jean Gabriel Tarde. It was presented by Carlos Pelta in the "2011 Meeting of the European Mathematical Psychology Group", celebrated in Paris.

The first law or law of Close contact (LCC) describes how individuals in close intimate contact with one another imitate each other´s behavior. The second law of imitation or imitation of superiors by inferiors (LSI) establishes people follow the model of high status in hopes their imitative behavior will get the rewards associated with being of a "superior" class. Tarde´s third law is the law of insertion (LOI): new acts and behaviors are superimposed on old ones and subsequently either reinforce or discourage previous customs. The following imitation rules are introduced: (1) Conf rule (Conformist rule) simulating the law of Close contact (LCC): if your behavior is different from that of the neighboring agent, copy its behavior; (2) Maxi rule (Maximization rule) simulates the law LSI and is so defined: if the neighbor agent gets higher payoffs, copy its behavior; (3) Fashion rule: copy the behavior with the highest frequency of appearance in your neighborhood (in case of equal frequency, copy at random); (4) Snob rule: copy the behavior with a lower frequency of appearance in your neighborhood (if the frequency of behavior appearance is the same, copy at random). Rules (3) and (4) simulate the law of insertion (LOI), alternating the copy of the latest choice made with the Fashion rule and the copy using the Snob rule in every round of the game. The agents have memory for these two rules for the 3 previous rounds of the game.

Once taken into account all these rules in a spatial prisoner´s dilemma, and combining all the possible values of b between 1 and 1.9, with an initial distribution of cooperators between 0.1 and 0.9, a memory M between 1 and 9 rounds for the rules (3) and (4) and changing the number N of agents and the number of rounds of the game, it is concluded that in our game the imitation rules by Tarde yield a preferential attractor and a low proportion of cooperating individuals. Although we have introduced two rules of stochastic nature (3) and (4), its effect is nullified by the proper mimetic dynamics, which means that they can not even be present in the attractor. Thus, agents attracted by non stochastic rules, and b values that increasingly are encouraging defection, are mass defined as defectors which find ways to maintain their payoffs as high as possible. But this circumstance supports Tarde´s law LSI because the imitation of the agents with higher payoffs (defectors) is majority also including the case with an initial rate of 0.9 cooperators receiving a payoff of 1 (defectors receive payoffs from 1.1 to 1.9). Besides our simulation verifies the law LOI, combining rules (1) and (2), because the most imitated behavior or Maximization behavior, makes via rule (1), the new behavior reinforced, discouraging the cooperative behavior of the agents with lesser payoffs.

The dynamics of group affiliation

Today we present in this blog the work by Nicholas Geard and Seth Bullock about the dynamics of group affiliation-see http://eprints.ecs.soton.ac.uk/21195/5/S0219525910002712.pdf-. Models about group formation are common in many social simulations. But models on group affiliation in which individuals can belong to multiple groups simultaneously are very infrequent. According to Geard and Bullock (art. cit., 2010, p. 501), some types of groups may be exclusive, that is, membership in one group precludes membership in other groups of that type and others are non-exclusive. Affiliation with a group involves the consum of time and energy being very important to determine the degree of commitment of the subjects and their degree of participation in other groups for studying the social evolution.

In pre-modern societies the affiliations were made in a series of concentric social circles from family to the country but in contemporary society all is more complex. In the "liquid society" (so called by Zygmunt), the bonds in choices of affiliation are very complex and fuzzy. Individual may belong to multiple groups simultaneously and Geard and Bullock design a model of affiliation to non-exclusive groups. Their simulation considers a network of n nodes and m undirected edges (art. cit., p. 507), representing individuals and the social ties between them. Each node i has a trait vector of dimension d, representing that individual´s location in social space, a list of affiliated groups and a time and energy capacity. Trait values are bounded between zero and one and are uniformly distributed. The social distance between two individuals is defined as the Euclidean distance between their trait vectors. Each group has a cost of time and energy associated with being a member, reducing the number of groups with which a node can be affiliated.

In the network, edges may be rewired either to nodes sharing a common state, or at random. Nodes may either initiate a new group, or be recruited to an existing group by one of their network neighbors. A node initiating a new group will always leave existing groups to maketime for the new group while a node being recruited to a new group will either leave existing groups or refuse the recruitment attempt, depending of the sociodemographic space.

For the simulations, the authors explored the circumstance where all memberships are exclusive, the population evolving to a "connected community structure" (art. cit., p. 509), that is, a type of continuing connectivity combined with the occasional initiation of novel groups. All groups had a cost of one but increasing cost above one had relevant effects on network structure, decreasing the level of community comparable to that of a random network. this trend suggests that as individuals belong to more groups, they are lees likely to become disconnected from the population, but have more opportunities to leave groups containing different to themselves. Obviously, less costly groups were maintained in the population in greater quantities than more costly groups but the mean size of the more costly groups remained constant as capacity of time and energy increased, while that of the less costly groups grown rapidly.

One interesting prediction is that less costly groups may find it easier to thrive, but that more costly groups may retain more diversity. We believe that the ideas surrounding the simulation by Geard and Bulloch is an interesting step forward for the modelization of the complex problem of the affiliation in social dynamics.

EMPG 2011: Forty years of the "European Mathematical Psychology Group"

(Telecom-Paris) (Photo by Carlos Pelta)

The "Meeting of the European Mathematical Psychology Group", which was held at the Telecom ParisTech, August 29-31, 2011, was a great success. Major credits go to Professor Olivier Hudry, the Meeting chair, who opened the Meeting with a few welcoming remarks. Next, Professor Marchant, the first plenary speaker, shared the latest developments about "Measurement theory with unary relations". H. Colonius and S. Rach have developed an approach based on the theory of Fechnerian Scaling for the measure of visual-auditory integration efficiency. Fechnerian Scaling deals with the computation of subjective distances from their pairwise discrimination probabilities. In the afternoon, L. Stefanutti spoke about knowledge structures extending the probabilistic framework to represent local independence among items in a probabilistic knowledge structure. Professors Alcalá-Quintana and García-Pérez introduced a model of indecision in perceptual detection tasks revealing strong order effects that vary in sign and magnitude in a systematic manner across observers. Besides, they used a probabilistic model of temporal-order perception to provide a common framework for synchrony judgments. Professor Shanteau described his experiments on memory-retrieval versus decision-making in repetition priming. Finally, Professors Albert and Hockemeyer analysed the very relevant contributions by Jean-Claude Falmage, the founder of the "European Mathematical Psychology Group", to Mathematical Psychology.

On Tuesday, Professor Raijmakers started the morning sessions with the oral presentation entitled "The application of latent Markov models in category learning". Latent Markov models allows for analysing multiple latent categorization strategies separately in a robust way. Next, Professor Pelta introduced a computational simulation in Social Psychology, adding to the spatial prisoner´s dilemma the three laws of imitation formulated by Jean-Gabriel Tarde in his book "The laws of imitation" (1890). Professor Thiel exposed how automata network models can simulate the halo effect in human attitudes, using a connectionist model on the Beckwith and Lehman multiattributes theory. In the afternoon, Jean-Claude Falmagne presented the idea of "Learning Spaces" and his colaborator Eric Cosyn introduced a very interesting practical application. Cosyn has extracted 350 items forming a learning space whose domain is the field of middle-school algebra. Professor Induráin tried to establish a common theory that relates the different mathematical properties that the concept of "mean" can have.

On Wednesday 31 August, Professor Choirat reviewed her work on separable representations in Mathematical Psychology and decision making. Finally, I would like to stress the oral presentation by Professor Doignon about representations of interval orders.

A post-conference edition of Meeting presentations should be available on the journal "Electronic Notes in Discrete Mathematics" perhaps during the first quarter of 2012.

We are very grateful, in first place, to the city of Paris, and, in a second place, to Professors Hudry, Lobstein, Charon and Choirat and Telecom ParisTech, for the organization of the Meeting.

2011 Meeting of the European Mathematical Psychology Group (Paris)

The "2011 Meeting of the European Mathematical Psychology Group" will be held at the TELECOM ParisTech, August 29-31, 2011 (http://www.telecom-paristech.fr/eng/home.html).

The conference is organized by Irène Charon (Tèlècom ParisTech), Olivier Hudry (Tèlècom ParisTech and CNRS), Antoine Lobstein (CNRS and Tèlècom ParisTech) and Hayette Soussou (Tèlècom ParisTech). The Program has been elaborated by Professor Hudry and the plenary speakers will be T. Marchant ("Measurement theory with unary relations"), L. Stefanutti ("When the correspondence between probabilistic and set representations of local independence becomes a requirement: constant odds models for probabilistic knowledge structures"), D. Albert and C. Hockemeyer ("JCF´s impact is not limited to the foundation of the EMPG"), M. Raijmakers ("The application of latent Markov models in category learning"), J.-C. Falmagne ("Learning spaces in real life. How the large size of actual learning spaces guides the development of the theory"), C. Choirat ("Separable representations in mathematical psychology and decision making") and A. Diederich ("Optimal time windows: Modeling multisensory integration in saccadic reaction times").

In the parallel sessions, the author of this blog (C. Pelta) will started the morning sessions on Tuesday 30 August (10:30 h.) with his oral presentation entitled "Spatial prisoner´s dilemma and laws of imitation in Social Psychology". I design a game based on the spatial prisoner introducing the three laws of social imitation defined by Gabriel Tarde in his book Les lois de l´imitation (1890). The French author described (a) the law of close contact (individuals in close intimate contact with one another imitate each other´s behavior), (b) the law of imitation of superiors by inferiors (people follow the model of high status in hopes their behavior will procure the rewards associated with the "superior" class) and (c) the law of insertion (new behaviors reinforce or discourage previous customs). I run a computational simulation in which the formation of little "clusters" of cooperators supports not only the laws of Tarde but also the ideas of Sutherland which explain the imitation of deviance behavior as a process of communication within intimate personal groups or "differential association".

I predict that the Meeting will be a great success and that the organization will be very succesful. The readers of this blog are cordially invited to participate. On September it will be published in this blog a summary exposing the main ideas of this event to celebrate in Paris. For more information, please, see the webpage content designed by Professor Olivier Hudry (http://www.infres.enst.fr/~hudry/EMPG/).

Computational Models of Human-Mate Choice and KAMA

Since classical article by Gale and Shapely (1962), several computational models about Human-Mate Choice have emerged. In this article, the authors developed a "match-making" algorithm for a population with an equal number of males and females. Kalick and Hamilton (1986) found a correlation in physical attractiveness among married couples. Kenrick et al. (2000) used dynamic social influence networks and concluded that males are inclined to take advantage of unrestricted relations whereas females prefer restricted relationships. Other models have been presented but in this article for the blog, we expose perhaps the most recent model. And for the author of this blog, perhaps the most interesting. It is adequately complex (it uses a vector of values to simulate the population-level effects of the modification over time of particular characteristics of individuals) and employs the mechanism of computational temperature for the simulation, that is, the amount of energy that people put into encountering and dating potential mates). Bob French and Elif Kus (2008) (see their article that was published in the journal Adaptive Behavior, http://leadserv.u-bourgogne.fr/files/publications/000261-kama-a-temperature-driven-model-of-mate-choice-using-dynamic-partner-representations.pdf) distinguish between "parallel" versus "serial" decision-making procedures. The male selects someone to ask out among a number of available alternatives ("parallel" decision process) and the female then accepts or declines his invitation immediately upon receiving it ("serial" decision process). KAMA, the computational model designed by French and Kus, implements the search of resources for a mate by a feedback-driven internal parameter called "temperature". In KAMA each agent has its own temperature that regulates its behavior. Temperature is a function of both an individual´s recent dating history and his/her age (French and Kus, 2008, p. 75), that is, a measure of the energy that one is willing to expend to find a partner. The higher the temperature, the more willing an individual is to explore for a mate; the lower the temperature, the less willing he/she is to do so. Also KAMA is a "stochastic model: essentially all choices are made probabilistically, on the basis of the individual´s temperature. The authors run a simulation (20 runs of the program) starting with 600 indviduals (half of them, females) whose ages vary randomly between 18 and 48. Both males and females maintain a list of all previously encountered individuals and the values of their characteristics, updated with each new encounter. After acceptance or refusal of a date, the temperature of the individuals involved is updated. The mechanisms of KAMA include "attractiveness" implying mate value. Characteristic preferences for the profiles are "kindness and understanding", "exciting personality", intelligence", "physical attractiveness", "good health", "adaptability", "creativity", "desire for childen", "College graduate", "good earning capacity", "good heredity", "good housekeeper" and "religious orientation". In addition to their preference profiles and characteristic profiles, all indviduals maintain a memory of all individuals they have previously encountered, along with the values of the characteristics of these individuals that they have discovered through encounters and dates with them.

To test KAMA, French and Kus drew on empirical data from the Eurostat. In KAMA, physical attractiveness decreased with age and wealth. On average, males´preference weighting for physical attractiveness was higher than the preference weight for females. The most surprising results were that when males and females had identical preference profiles and identical temperature curves, there was a marked male-female hazard-rate shift. Why does the fact that males ask women out and women accept or refuse lead to this difference in hazard rates? The asymmetry in the males-ask/females-decide custom produce this difference in hazard rates. When women can ask men out, this asymmetry disappears and, all other things being equal, the male-female hazard-rate shift disappears.

More sophisticated versions of this model are necessary but we think that KAMA incorporates novel features like the notion of agents with indidualized preferences or the idea of computational temperature which controls the focus of decision making. Undoubtely, KAMA is a very functional and complete model for the Human-Mate Choice.

(Photo: Bob French).

Models of minority opinion spreading

Maxi San Miguel (Physics of Complex Systems, University of Balearic Islands) has won the Medal of the Royal Society of Physics in Spain. Congratulations for this excellent researcher. His studies linking Physics to Social Dynamics are very interesting. And so, Wio, Toral, Tessone, Amengual, San Miguel (2004), have exposed several neighborhood models of minority opinion spreading the idea which we analyze in this article. According to the authors, the neighborhood models are locally defined neighborhood cells systems with complete connectedness. Neighborhood cells change shape and size during evolution. The question is How an initially minority opinion can become majority? Obviously, there are several theoretic models like Galam Model. Galam (2002) says that social inertia is a conservative response to the risk of a change maintaining social status quo. Let be a binary opinion and initially there is a minority against social reform. Cells are defined only by their size. A tie in the voting is a "No" for social reform. When all the agents in the cell adopt an opinion, agents join a meeting cell randomly selected. Decision rule is applied in all the cells. Agents randomly redistributed in the meeting cells carry their adopted opinion. Applying a mean-field analysis, there is a threshold value of initial minority supporters such that the minority opinion finally becomes majority. There is an asymmetric unstable fixed point or "faith point". Time to reach consensus is fast and system-size independent. In this model, individuals are fixed at the sites of a regular lattice and meeting cells are locally defined by a tessellation of the lattice. Consensus is always reached in finite systems in a finite number of steps. In an infinite system the initial minority opinion wins regardless the amount of initial supporters. Why? Because a critical size for an initial local domain of minority supporters exists: a domain of overcritical size always exists in a large enough population. Neighborhood models describe a more efficient spreading of minority opinion, but spreading takes a much longer time.

Social groups and chaotic state transitions: homage to Walter J. Freeman III

In this article, we pay homage to one of the most prominent neuroscientists ever: Walter Jackson Freeman III. He has designed a perspective called Nonlinear Neurodynamics of the brain that, perhaps is the most advanced and veridical approximation to the study of the brain dynamics. More interesting for this blog is the connection between his neurophysiological discoveries and its applications to the social dynamics o formation of social groups (see his book, "Societies of Brains", 1995). According to Freeman (1995), the cerebral cortex switches abruptly from one basin of attraction to another, each transition involving learning. Therefore, each brain creates its own trajectory which is not directly accessible by any other brain. The question is: how can several brains be shaped by learning so as to form cooperative groups for survival and reproduction? Large numbers of neurons follow chaotics dynamics expressing global state transitions (sleep to waking, etc.) and one class of state transitions in brains provides for the formation of social groups. Brains process meaning. But this intentional mechanism implies, in a certain sense, the isolation of each brain. With respect to energy and information each brain is an open system but with respect to meaning it is a closed system. However Nature has evolved powerful methods for the social learning and social cooperation. The discovery of the means for inducing these forms of learning can be understood as a chaotic state transition in brain dynamics like, for instance, the rapid adaptation of young adults for their new roles in state transitions from child to adulthood.

Maja J. Mataric and social robots

In this article we expose some interesting ideas by Maja J. Mataric (University of Southern California) about the design of social robots (see Mataric, 2002 in Encyclopedia of cognitive science).

Building sociable robots includes many facets, like imitation, social learning and emotion. We can design social robots applying ideas from developmental psychology, for example and so looking for inspiration in Neuroscience. Mataric (1992) described the work with TOTO, a mobile robot being able to represent landmarks in the environment. TOTO was representative of an Artificial Intelligence interpretation of the organization of the rat hippocampus. As an alternative, Nicolescu and Mataric (2002) designed a hierarchical behavior-based architecture enabling behaviors to represent more abstract concepts. Here representations are stored in a distribuited fashion. The same perspective on generating behavior has been successful with groups of robots. This area is known as "swarm robotics". Truly coordinating a set of robots is complicated problem. Multi-robot coordination involves communication and selection action, between several tasks. In 1995 Mataric worked with the NERD HERD, a group of 20 autonomous mobile robots with limited sensing and computational abilities. Each robot was programmed with a small set of behaviors: homing, wandering, following, aggregation and dispersion. The basis behaviors were designed to conserve energy by minimizing interference between robots. Mataric (1997) described the problem of learning social rules in order to maximize energy. Robots acting within a social setting have additional sources of information: observation of a peer performing a successful action, etc. Models of people´s natural social interactions are relevant for robots in human environments. For humanoid robots this can take the form of learning natural human skills. Mataric is an excellent searcher looking for solutions which imply the design of social robots.

Affect Control Theory and Social Models of Human-Computer Interaction

One of the most interesting models of emotion is Affect Control Theory (ACT) (Heise, 1987). ACT expresses how social events are construed positing a relation between cognition and emotion. In contrast to psychological theories of social cognition, ACT emphasizes that meanings are culturally shared and deviations from meanings generate arousal that triggers re-appraisals. In this brief article and following to Troyer (2004), an outstanding specialist, I expose some aspects of the theory and how Lisa Troyer applies its concepts to model Human-Computer Interaction (HCI).

ACT looks at individuals as agents seeking consistency across interactions. Humans are categorized into roles with shared expectations regarding actions appropriate for the role. Actions in response to one another are markers of social events. Social events are formed by actors who assume identities, behaviors of actors and objects to whom the actions are directed. For instance, a policeman helping citizen, whose linguistic structure would be "Policeman Helps Citizen". Its meaning is defined in three dimensions: Evaluation (goodness), Potency (powerfulness) and Activity (liveliness). Ratings of each element are called "fundamental sentiments" (for it the word affect). Humans have culturally shared fundamental sentiments and expectations. So, we expect good policemen to behave in right ways. When elements combine in an event, emotion signals the corespondence between the meanings we expect and the actual meanings evoked. Smith-Lovin (1987) introduced impression-formation equations combining ratings of elements to estimate new ratings for inputs combined in events. These equations predict how meanings shift as interaction evolves. The sum of the squared differences between fundamental sentiments of the elements and impressions from the event generates the perceived likelihood of an event.

ACT includes a database of ratings for thousand of elements and modifiers (emotion labels for roles, for instance, "pleasant policeman"). Using the equations and database, ACT predicts the basis for expectations in subsequent interaction. The models and database are combined in the software INTERACT. With this software researchers simulate events and generate testable predictions regarding sequences and redefinitions of events. ACT has focused on Human-Human interaction but Troyer (2004) demonstrates how can be used to model Human-Computer interaction. ACT does not require that computers exhibit emotions, but only that they be able to reason about them, that is, metareasoning.

Troyer ("Affect control theory as a foundation for socially intelligent systems", 2004) designed an experiment with 15 subjects, providing them independent ratings for elements of Human-Computer interaction: "Computer", "Run Analysis", "Provide Output", "Freezes" and "Runtime Error". Correspondent social concepts were "Academic", "Ask about Something", "Educate", "Beg" and "Laughs At". Troyer substituted the correspondent social concepts for HCI terms to simulate events representing Human-Human interaction analogs of HCIs. The simulations explored how meanings shifted when a computer initially behaved as expected, producing unexpected behavior. The simulations showed how different events produced different definitions of the actor eliciting the behavior (grouch/delinquent) and the responses to that actor (scold/avoid). Besides, ACT predicted the emotions of the object receiving the behavior.

ACT may provide an architecture for designing socially intelligent systems.

Physarum machines

A Physarum machine is a programmable amorphous biological computer experimentally implemented in the state of slime mould Physarum polycephalum. It comprises an amorphous yellowish mass with networks of protoplasmic tubes, programmed by spatial configurations of attracting and repelling gradients. It feeds on bacteria, spores and other microbial creatures. When foraging for its food the plasmodium propagates towards sources of food particles, surrounds them, secretes enzymes and digests the food. The plasmodium is considered as a parallel computing similar to existing massive parallel reaction diffusion chemical processors. Adamatzky (University of Bristol) and collaborators demonstrate how to create experimental Physarum machines for general purpose computation: plasmodium can implement the Kolmogorov-Uspensky (KUM) machine, a mathematical machine in which the storage structure is an irregular graph. KUM is defined on a labeled undirected graph with bounded degrees of nodes and bounded number of labels. KUM executes several operations on its storage structure: SELECT an active node (that is, occupied by an active zone) in the storage graph; SPECIFY the node´s neighborhood; MODIFY the active zone by ADDING a new node with the pair of edges, connecting the new node with the active node; DELETE a node with a pair of incident edges; ADD/DELETE the edge between the nodes. A program for KUM establishes how to REPLACE the neighborhood of an active node with a new neighborhood, depending on the labels of edges connected to the active node and the labels of the nodes placed in proximity of the active node. In Physarun machine, a node of the storage structure is represented by a source of nutrients, an edge connecting two nodes is a protoplasmic tubes linkink two sources of nutrients corresponding to the nodes. Finally, an active node is domain of space (which may include nutrient sources) occupied by a propagating pseudopodium. The computation is implemented by several active zones. It uses distributed local sensory behaviours, approximating phenomena observed in Physarum, like foraging for food stimuli, amoebic movement, network formation, network minimisation, surface area minimisation, shuttle streaming, spatially distributed oscillations or oscillation phase shifting. The emergent plasmodium behaviours are represented taking a multi-agent approach or based upon particles. In fact, movement and internal oscillations of the plasmodium reflect the collective behaviour of the particles population. The movement of agents correspons to the flux of sol within the plasmodium. Cohesion of the plasmodium arises due to the agent-agent interactions and movement of the plasmodium is generated by coupling the emergent mass behaviours with chemoattraction to local food source stimuli. Agents both secrete and sense approximations of chemical trails being the population represented on a two-dimensional discrete map. The strength of the projected food sources can be adjusted using parameters and when the plasmodium engulfs a food source the stimulus for diffusion is reduced by the encapsulation. The diffusion gradient corresponds to the quality of the nutrient and substrate of the plasmodium´s environment, and differences in the stimulus strength, stimulus area, affect both the steepness, and propagation distance of the diffusion gradient and affect the growth patterns of the virtual plasmodium.

The Physarum machine by Adamatzky is a very interesting example of a green computer, showing the necessity for simulating simple behaviours, in first place, as motivation for developing more complex simulations. It will be an excellent source of ideas for anyone who is inspired by emerging non-silicon computers.

Pandas: Pandaemonium-Controlled Animats

(Artificial Life, vol. 16, 1, Winter 2010, The MIT Press; cover by Philip Beesley´s Epithelium, an installation at the Siegel Gallery at Pratt Institute of Design in Brooklyn, N.Y., 2008).

On this month I would like to expose an interesting model of cultural transmission developed by Chris Marriott, James Parker and Jörg Denzinger at University of Calgary. It is called PANDAS and simulates the effects of an imitation mechanism on a population of animats (artificial animals) capable of individual ontogenetic learning.

Let be a world inhabited by animats and consisting of a discrete grid that contains besides the following objects: food, water, cave and tree. The world obeys several rules of evolution: (a) no object can occupy the same cell as a tree; (b) all objects are stationary except pands; (c) water, caves, and trees are static and (d) food is depleted when used, and grows over time. For the pandas, we have these rules: (1) pandas can move one cell in one of eight directions (N,NE,E,SE,S,SW,W,NW); (2) pandas obtain food energy, water energy, rest energy and have a health parameter and, in every round, lose a fixed amount of food energy, water energy and rest energy; (3) pandas die if any parameter in (2), drops below cero; (4) a dead panda is removed from the grid and replaced with food; (5) a panda can only mate with another mating panda; (6) a panda can create a new panda by mating or spawning but a panda can only mate or spawn when it has an excess of all energy types; (7) pandas can only interact with objects in the same cell; (8) a panda can fight with another panda, reducing the target panda´s health and (9) a panda´s health automatically recovers, at an energy cost.

A panda has only a one-cell perception range and has a limited internal sense allowing it to monitor its energy levels. All perception is done using input daemons. Daemons are similar to nodes in a network playing a specific role. The input daemons that form the input layer of the panda include 9 daemons for each object in the environment (see food, see water, see cave, see tree). There is one daemon for each of the cells in the perceptual range of the panda. The input layer contains the following set of daemons (food energy, water energy, rest energy, hungry, thirsty, tired). The cognitive cycle of a panda begins with taking input from the environment and activating input daemons, and ends with an action selected. When an action has been selected, it is executed in the system and the cycle begins with the new situation.

The model is very interesting because pandas can engage in three types of adaptation: the genome of a panda encodes both "physical properties" (the panda interacts with its environment) and "mental properties" (the panda maintains its internal organization). To study the effects of the imitation on the populations of pandas, it is considered a group of pandas without imitation drive and unable to perceive directly and a second group of imitating pandas. The authors use the life span of the pandas as an indicator of their success. To do this, they introduce a parameter called the elder age, that is, an arbitrarily selected age such that when a panda reaches this age before dying, is considered successful and the value is used for the measurement of frequency. According to Marriott, Parker and Denzinger, the data of the experiments using this model, show that the median frequencies for imitating pandas are hogher than the frequencies for non-imitating pandas. Obviously, the imitating pandas have the tendency to group together, being particularly strong in newborn pandas. The model supports that mechanisms of cultural transmission can increase the frequency of success in a population but the authors look for extending it to the more sophisticated mechanisms of the "true imitation" because they think that the attribution of higher-order intentionality to artificial agents is something pending nowadays.

Gaze following and mirror neurons

Gaze following is a basic component of the human social interaction and it is a type of attention sharing behaviors. It is also present in a number of other species (for instance, apes) and seems necessary for designing social robots. It can be defined as the ability to look where somebody else is looking. Triesch, Jasso and Deák (2007) have formulated a computational model of gaze following by means of ideas related to the behavior of mirror neurons. The authors emphasize the role of learning processes by means of the interaction with the social environment. In fact, gaze following can be linked to imitation. The link between gaze following and imitation also implies the similarity between the neural basis of gaze following and the neural basis of other imitative behaviors. Triesch and collaborators develop a model that share properties with mirror neurons, that is, neurons implicated in imitation and originally founded in macaque area F5 by Rizzolatti and his team in Parma University. The model predicts the existence of a new class of mirror neurons for looking behaviors that has not been observed experimentally. In this model, an infant and a caregiver interact with a number of visually salient objects. During the process, the infant learns to predict the locations of salient objects based on the looking behavior of the caregiver. There are periods when the caregiver is present and periods when the infant is alone with the objects. When the caregiver is present, the infant and caregiver are in fixed locations facing each other with a separation between them. At any time a random number of objects will be present. Habituation decreases the perceived saliency of an object.

The infant model learns through a reinforcement learning scheme. The learning process tries to optimize the infant´s policy, that is, the way the actor maps sensory states onto different gaze shifts in order to maximize the long-term reward obtained by the infant. Reward is obtained as the saliency of the position to which attention is directed after a gaze shift has been made. At each time step, the caregiver looks at the most salient object, where saliency is mediated by the same habituation mechanism as in the infant´s visual system. The model neurons in the pre-motor layer share many characteristics with classical mirror neurons. A unit in this layer will be active during the execution of a gaze shift to a certain location in space. This is because the probability of performing such a gaze shift is related to the activation of the unit. The units in the layer will be active when the infant observes the caregiver looking in the corresponding direction. Clearly, the neurons in the pre-motor layer can be viewed as mirror neurons because the combination of being active during execution and observation of a motor act is the defining characteristic of mirror neurons.

Following to the authors, this model can be considered a simple associative learning account of a response facilitation but also has implications for the question of whether mirror neurons are innate or whether they acquire their properties through a learning process. For the mirror neurons concerned with grasping, they find plausible that there are situations where observing an agent grasp an object may predict a reward if the same action is attempted. Such situations may be sufficient for the emergence of mirror neurons for grasping. The reviewed model predicts a very close connection between mirror neurons and imitative behaviors.

Dynamic Neural Fields and cooperative robots

Dynamic Neural Fields formalize how neural populations represent the continuous dimensions characterizing movements, perceptual features and cognitive decisions of agents. Neural fields evolve dynamically generating elementary forms of cognition. Many of social activities are based on the ability of individuals to predict the consequences of other´s behavior. We have to interpret actions of our partners in collaborative works.

Erlhagen et al. (2007) try to understand motor intentions for building artificial agents using Dynamic Neural Fields. The authors consider that action understanding relies on the notion that the observer uses its motoric abilities to replicate the observed actions and its effects. They are inspired in activity patterns of mirror neurons in prefrontal cortex that postulate a chain between neurons coding motor acts. Depending on the specific chain of mirror neurons activated by contextual and external cues, the observer will predict (in a probable manner) what the observed agent is going to do. Erlhagen and collaborators represent the activity of neural populations encoding different motor acts and goals by means of Dynamic Neural Fields. The synaptic links between any two populations in the network is established using a Hebbian learning dynamics.

Let think in autonomous robots which interact in the context of a join construction task and let be a simple reaching-grasping-placing scenario. An observing robot R1 has to select a complementary action sequence depending of the inferred action goal of the other robot R2, or partner robot. So, R2 may grasp an object to place it in front of R1 with the intention to hand it over. Neural populations in the action observation layer and the action simulation layer encode motor primitives such as grasping. Such neural populations in the goal layer are associated with the respective chains in the action simulation layer. To model the dynamics of the different neural populations is used a discrete version of a dynamic neural field. Each dynamic field represents a population of 2N neurons which diverge into an excitatory and an inhibitory subpopulation, each of dimension N. The activation of an excitatory and an inhibitory neuron i at time t is governed by a coupled system of differential equations. The firing rate and the shunting term for the excitation, are non-linear functions of sigmoid shape and the interaction strength between any two neurons within the subpopulation is established by fixed synaptic weight functions which dicrease as a function of the distance between the neurons. A Hebbian learning rule for increasing the synaptic efficacy between presynaptic and postsynaptic neurons is given.

For establishing the chains, the authors employ a learning by observation paradigm in which a teacher demonstrates the sequences, each composed of motor primitives. Once the neural population becomes active, the activity propagates to all synaptically coupled populations. But only a population defining a particular chain will reach a suprathreshold activation level.

The simulation of Erlhagen et al. 2007, shows that the neural representations implementing mechanisms like motor simulation and cue integration may emerge as the result of real-time interactions of local neural populations. But, more important is that they speculate that learning by imitation takes place in two steps: first, the links between chain elements are established allowing a fluent execution of particular action sequences. In a second place, similar action sequences may have a different outcome and then, the focus shits towards the links between the goal and the contextual cues. These consequences will be crucial in the cooperative robotics domain.

REFERENCE
(http://www.mct.uminho.pt/erlhagen/pdfs/ErlhagenEtAl_ICDL07_0004.pdf)

Mirror neurons and synchronization between robots

The discovery of "mirror neurons" in the ventral premotor cortex of the macaque monkey by Rizzolatti and colaborators has generated a genuine impact in Neuroscience. In humans it is impossible to registry the neurophysiological activation of simple neurons but disregarding criticisms (see Alison Gopnik-http://www.slate.com/id/2165123/pagenum/all/-), the influence in many fields of the knowledge (study of social relations, robotics, programming, etc.) is enormous. Privileged witness is the recent work of Barakova, Lourens and Yamaguchi. Barakova and Lourens (2008) design a setup for synchronization and turn-taking behaviour in robots. For it, they connect some aspects of the neuroanatomy of the mirror system in humans with an oscillatory dynamics for neural networks.

In brief, the frontal motor areas receive sensory input from the parietal lobe. Another area is situated in the rostral part of the inferior parietal lobule. Both regions form the mirror neuron system. Besides, the posterior sector of the superior temporal sulcus form a core circuit for imitation. Modelling of the superior temporal sulcus area can be reduced to the influence of the inhibitory neurons, projecting the sensory signals to the inferior parietal lobule, area which is associated with multisensory integration. Each robot has 8 range sensors projecting to the sensory integration area that resembles the functionality of the joined temporal sulcus-inferior parietal areas. The two wheels of the robot project to the sensorimotor integration area, resembling the function of the ventral premotor cortex. Self-organization of rhythmic activity of this system can be simulated by means of endogenous oscillators an so, the change of rate of the phase with the time, is the cycle of the limit cycle oscilation, being the phase periodic over the range.

The experimental setting conceived by Barakova and Lourens presents, in the first place, robots that are taking the role of the follower, in order to establish "mirroring" couplings between the nets that simulate the inferior parietal lobule and premotor ventral areas. Hebbian connections between these nets are modelled such that the interaction behaviour is reflected by the average activation values of unikts over a certain time interval. So, the robot playing the role as a follower, tends to synchronize its motion direction with the motion direction of the leading robot. In the second place, the experiment shows the emergence of turn taking between the robots. The role of the robot, being follower or leader depends on which robot is within the visual field of its partner. The emergent turn taking is expressed by symmetry breaking process after a period of synchronization: the leading robot can become a follower and later the lead can be taken over by it.

Inspired by the mirror system in human beings, Barakova and Lourens simulate very interesting interaction behaviours of following and turn taking such that the mirroring functionality is obtained by means of the selforganization of synchronized neural firing in two robots that share perceptual space. No doubt that many extensions of this work remain to develop and promise great advances in the area of robotics.

The e-pucks and social cooperation

The e-pucks are mobile robots developed at the École Polytechnique Fédérale de Lausanne (EPFL). The designers (see Mondada et al., The e-puck, a Robot Designed for Education in Engineering, 2009) looked for the miniaturization of a complex system combining desktop size and flexibility. These robots have sensors in different modalities (distances to objects by means of eight infrared proximity sensors, accelerometer, color camera, microphones), actuators with different actions on the environment, wired and wireless communication devices and two types of processors (general purpose and DSP). Although they were conceived for education in Engineering, they have demonstrated to be very useful for experimentation in Artificial Intelligence.

Social learning is the capability of an organism to learn by observing the behavior of a conspecific. Evolutionary robotics is a methodological tool to design robots´controllers. Recently, Miglino, Ponticorvo and Donetto (2008)

(http://www.nac.unina.it/images/pub/cooperationincorvid.pdf)

have used an evolutionary algorithm for the neural control of e-puck robots which imitate the cooperation in corvids to obtain a reward (food), which is clearly visible, but not directly reachable. The dyad gets the reward if the two tips of a string are pulled at the same time.

Two e-puck robots are situated in an environment consisting of a square arena and of a corridor both surrounded by walls. The authors use an evolutionary algorithm to set the weights of the robots´neural controller. The initial population consists of 100 randomly generated genotypes that encode the connection weights of 100 corresponding neural networks. Each genotype is translated into 2 identical neural controllers which are embodied in 2 e-pucks. The 20 best genotypes of each generation are reproduced by generating 5 copies each, with 2% of their bits replaced with a new randomly selected value. The evolutionary process is iterated 1000 times and the experiment is replicated 20 times each consisting of 4 trials with 4 different starting positions in the corners of the room. Cooperation between e-pucks is regulated by social interaction with communication as a medium. The emergence of communication leads to a coordinated cooperation behavior similar to cooperation observed in corvids.

Nowadays artificial evolution is employed to build neural mechanisms that control the behavior of learning robots. Mechanisms for social learning in organisms can be successfully simulated in an integrated neural network architecture. E-pucks are an excellent tool for simulating social learning in organisms requiring cognitive mechanisms. But future work is needed, in particular to allow the robots to autonomously decide when to use social learning strategies or individual strategies.

Goran Trajkovski on imitation to modeling agents societies

In this brief review we analyse some ideas about simulation of imitation by means of artificial agents. Dr. Trajkovski is Director of Cognitive Agency and Robotics Lab in Towson University (USA). He is specialized in Cognitive Engineering and has published books like "An imitation-based approach to modeling homogenous agents and societies" or the recents "Handbook of Research on Agent-Based Societies: Social and Cultural Interactions" or "Handbook of Research on Computational Arts and Creative Informatics". Trajkovski introduces agents capable of performing four elementary actions (forward, backward, left, and right) and of noticing 10 different percepts. Each agent is equipped with one food sensor and has one hunger drive. When the hunger drive is activated for the first time, the agent performs random walk during which expectancies are stored in the associative memory. When food is sensed, expectancy emotional context is set to a positive value. Every time in the future the hunger drive is activated, the agent uses the context values of the expectancies to direct its actions. It chooses the action that will lead to expectancy with maximum context value.

Agents inhabit a two-dimensional world surrounded by walls and populated with obstacles. Sensing another agents takes the agent into its imitating mode. The agents have begun to inhabit the environment at different times. They are also being born in different places in the environment. While inhabiting the environment, they explore different portions of the environment that may be very different, or perceptually similar. In environments inhabited by heterogeneous agents, the fundamental problem is the problem of interagent communication. According to Trajkovski, an example of interagent communication is the phenomenon of multilingual agents that can serve as translators. The author proposes an enactivist (Varela, Thompson and Rosch, 1991) representation model, based on the treatment of agents as dynamical systems. The agent during the interaction with the environment generates its inputs and makes a mapping from the continuous domain of the inputs to the discrete domain of the percepts. The sequence of these percepts would reflect the structure of the environment. The basic idea is to divide up the set of possible states into a finite number of pieces and keep track of which piece the state of the system lays in at every iteration. Each piece is associated with a symbol, and in this way the evolution of the system is described by an infinite sequence of symbols. The agent generates symbols or percepts.

Trajkovski shows that imitation is far from a trivial phenomenon and that humans are wired for imitation by means of the research in multi-agent systems. He gives a solid and fertile attempt to establish a mechanism of interagent communication in the multi-agent environment using classical algebraic theories and fuzzy algebraic structures. His contributions are very relevant for the social cognition, mixing studies about animal intelligence (Thorndike) with studies about imitation in humans.

A New Computational Model of Social Learning

We describe a very interesting computational model of social-learning mechanisms proposed by Lopes, Melo, Kenward and Santos-Victor (2009). The authors distinguish between imitation ("adhere to inferred intention, replicate observed actions and effect") and emulation ("replicate observed effect") taking into account the agent´s preferences for different actions and the information available from the demonstration. There is a module addressing the baseline preferences of the agent, evaluating actions in terms of energy consumption. With this module is associated the utility function Qb ranking possible action sequences according to their overall energy consumption. A utility function Qe evaluates the actions in terms of their probability of reproducing the observed result/effect. For the intention replicating module, the utility function Qi assumes the demonstrator is goal-oriented. The module operates by considering all the possible goals in the current system, calculating for each one the relative probability that it would give rise to the demonstrated behaviour, and choosing the one that maximises the probability. In cases that are equally likely to produce the observed demonstration, goals with tied probability are ranked randomly which leads to stochasticity in the final behaviour.

The model of Lopes and collaborators can replicate the tendency to interpret and reproduce observed actions in terms of the inferred goals of the action in line with the experiments of Malinda Carpenter and colleagues (2005): a demonstrator moved a toy mouse across a table from one point to another; in one condition, the final point of the move was inside a little house, and in the other condition, no house was present: infants showed a much grater tendency to replicate the specific mouse moving action observed when there was no house to move the mouse into. The results of the simulation reproduce the findings of Carpenter et al. (2005), confirming the standard interpretation of the experiment: the infant infers what the demonstrator´s intention was.

Other simulation replicates the famous experiment of Meltzoff (1988) in which 14-month olds were exposed to a demonstrator who performed unusual actions on objects (there was a box with a panel that lit up when the demonstrator touched it with his forehead, and most infants copied the use of the forehead rather than using their hand): the infants reproduced the actions with a delay of a week. According to Lopes and collaborators, the simulation confirms the imitation in terms of the inferred intention and of sensitivity to the constraints on the demonstrator.

To investigate what happens when the learner does not have complete knowledge of the world dynamics, they model a type of experiment based on Horner and Whiten (2005) experiment in which presented preschoolers and chimpanzees with two identical boxes, one opaque and one transparent. The demonstration consisted of inserting a stick into a hole on the front of the box, with the latter step generating a reward. The insertion of the stick into the top hole was unnecessary in order to obtain the reward, but the causal physical relations were only visible with the transparent box. The results showed that 3 and 4-year-old children imitated both actions no matter whether they had observed and were tested on the transparent or opaque box, but chimpanzees were more able to switch between emulation and imitation after having observed demonstrations with a transparent box and reduced tendency to insert the stick into the ineffective hole.

The simulation designed by Lopes, Melo, Kenward and Santos-Victor replicate the results from both children and chimps when the weight of the intention replicating module was increased, confirming that the difference occurs because chimps are primarily motivated to select the most efficient method they know to achieve the end effect.

Following the taxonomy proposed by Call and Carpenter (2002), Lopes et al. (2009) build a unifying mathematical model of types of social influence on behaviour, mainly imitation and emulation, concluding that a switch between imitation and emulation might be triggered by changing the value (to the learning) of the social interaction or of the effect. So, the greater utilization of imitation by children might be explained by a stronger focus on others´intentions, mediated by social cues.

Imitation, Maslow´s Pyramid and Multi-Agent Systems

We review an interesting article published by Le Guen Herve and Moga Sorin in Proceedings of International Joint Conference on Neural Networks (2009).

The authors study the influence of imitation within a population of artificial agents following Maslow´s Pyramid of needs. Imitation is the possible origin of communication and social learning. The imitative abilities include many types like facial imitation, perception by the infant that he or she is being imitated and empathy. The interest in empathy induces the study of emotions which Herve and Sorin link to the analysis of social needs as modeled by the Maslow´s Pyramid of motivations. Although the influence of each particular need varies from one person to another the principle is there are two main classes of needs. The most important needs are called primary needs (eating, breathing, etc.) while the second class of needs constitute the social aspect of human motivation and includes the need for esteem. This need is bidirectional and has been interpreted by the authors as the need to imitate and the need to be imitated. As the fact of being imitated is perceived by the agent concerned and as imitation is likely to generate empathic satisfaction, the authors model them as expected rewards. Based on a model of an autonomous robot with goal-oriented navigation and imitation capabilities, the robot goals are derived from internal variables that have to be maintained in a comfort area. The values of these variables decrease in time. The robot population´s task is to explore an unknown environment and to localize sources corresponding to its needs. Its survival will depend upon the satisfaction of these needs by discovering the different locations of the sources. The behaviors are obstacle avoidance, goal-oriented navigation, imitation and exploration. The emotional signature expresses the current state and the expected state of the agent: an agent in its comfort area displays a neutral signature whereas an internal variable below a certain threshold induces pain. That pain will cause a potential empathic response that is likely to incite another agent to move toward a known source. Besides a motivated agent is likely to provoke an attractive empathic response: keeping a motivated agent in its own field of perception becomes a source of motivation. It leads to the selection of the imitation target according to its motivational state.

The results of the simulation with Maslow agents showed an enhancement of the global survival rate even with a very small population wherein communications are not frequent.

The authors have proposed a holistic approach to the implementation of imitation in autonomous agents.

We recommend this excellent article because its ideas permit to build a simple and scalable model of agent useful in applications such as networks or swarm piloting.