Adaptive Behavior recent issues

A Computational Model of Social-Learning Mechanisms

Lopes, M., Melo, F. S., Kenward, B., Santos-Victor, J. — Thu, 19 Nov 2009 04:47:11 PST

In this article we propose a computational model that describes how observed behavior can influence an observer’s own behavior, including the acquisition of new task descriptions. The sources of influence on our model’s behavior are: beliefs about the world’s possible states and actions causing transitions between them; baseline preferences for certain actions; a variable tendency to infer and share goals in observed behavior; and a variable tendency to act efficiently to reach rewarding states. Acting on these premises, our model is able to replicate key empirical studies of social learning in children and chimpanzees. We demonstrate how a simple artificial system can account for a variety of biological social transfer phenomena, such as goal-inference and over-imitation, by taking into account action constraints and incomplete knowledge about the world dynamics.

Incremental Learning and Memory Consolidation of Whole Body Human Motion Primitives

Kulic, D., Nakamura, Y. — Thu, 19 Nov 2009 04:47:11 PST

The ability to learn during continuous and on-line observation would be advantageous for humanoid robots, as it would enable them to learn during co-location and interaction in the human environment. However, when motions are being learned and clustered on-line, there is a trade-off between classification accuracy and the number of training examples, resulting in potential misclassifications both at the motion and hierarchy formation level. This article presents an approach enabling fast on-line incremental learning, combined with an incremental memory consolidation process correcting initial misclassifications and errors in organization, to improve the stability and accuracy of the learned motions, analogous to the memory consolidation process following motor learning observed in humans. Following initial organization, motions are randomly selected for reclassification, at both low and high levels of the hierarchy. If a better reclassification is found, the knowledge structure is reorganized to comply. The approach is validated during incremental acquisition of a motion database containing a variety of full body motions.¹

Pattern-Oriented Modeling of Commons Dilemma Experiments

Janssen, M. A., Radtke, N. P., Lee, A. — Thu, 19 Nov 2009 04:47:11 PST

A major challenge in the development of computational models of collective behavior is the empirical validation. Experimental data from a spatially explicit dynamic commons dilemma experiment is used to empirically ground an agent-based model. Three distinct patterns are identified in the data. Two naïve models, random walk and greedy agents, do not produce data that match the patterns. A more comprehensive model is presented that explains how participants make movement and harvest decisions. Using pattern-oriented modeling the parameter space is explored to identify the parameter combinations that meet the three identified patterns. Less than 0.1% of the parameter combinations meet all the patterns. These parameter settings were used to successfully predict the patterns of a new set of experiments.

Partner Search Heuristics in the Lab: Stability of Matchings Under Various Preference Structures

Eriksson, K., Strimling, P. — Thu, 19 Nov 2009 04:47:11 PST

When agents search for partners, the outcome is a matching. K. Eriksson and O. Häggström (2008) defined a measure of instability of matchings and proved that under a certain partner search heuristic, outcomes are likely to have low instability. They also showed that with regards to stability, the preference structure known as common preferences lie somewhere in between the extreme cases of homotypic and antithetical preferences. Following up on this theoretical work, we let human subjects search for a good partner in a computer game where preferences were set to be either common, homotypic, or antithetical. We find that total search effort and instability of the outcome vary in the predicted ways with the preference structure and the number of agents. A set of simulations show that these results are consistent with a model where agents use a simple search heuristic with a slight possibility of error.

Levels and Types of Action Selection: The Action Selection Soup

Ozturk, P. — Thu, 19 Nov 2009 04:47:11 PST

Action selection (AS) is defined as the process where an action is selected among a number of alternatives. This definition, however, does not sufficiently describe what an action is. What is the unit of selection in the first place? We maintain that the artificial intelligence (AI) accounts of AS typically mix and merge two AS situations that indeed are qualitatively different. Most of the accounts actually deal only with one type of AS but purport to cover both types of AS. We propose three dimensions along which the commonalities and the differences between various AS accounts can be analyzed, and use these for a preliminary conceptualization of what we call a two-system action selection account. In particular, we identify two qualitatively different AS situations whose architectures, we suggest, can be designed inspired by neuroscience models of the basal ganglia (BG) and the cerebellum, respectively.

Thanks to Reviewers

Thu, 19 Nov 2009 04:47:11 PST

Editorial: Agency in Natural and Artificial Systems

Rohde, M., Ikegami, T. — Wed, 23 Sep 2009 02:33:53 PDT

Defining Agency: Individuality, Normativity, Asymmetry, and Spatio-temporality in Action

Barandiaran, X. E., Di Paolo, E., Rohde, M. — Wed, 23 Sep 2009 02:33:53 PDT

The concept of agency is of crucial importance in cognitive science and artificial intelligence, and it is often used as an intuitive and rather uncontroversial term, in contrast to more abstract and theoretically heavily weighted terms such as intentionality , rationality, or mind. However, most of the available definitions of agency are too loose or unspecific to allow for a progressive scientific research program. They implicitly and unproblematically assume the features that characterize agents, thus obscuring the full potential and challenge of modeling agency. We identify three conditions that a system must meet in order to be considered as a genuine agent: (a) a system must define its own individuality, (b) it must be the active source of activity in its environment (interactional asymmetry), and (c) it must regulate this activity in relation to certain norms (normativity). We find that even minimal forms of proto-cellular systems can already provide a paradigmatic example of genuine agency. By abstracting away some specific details of minimal models of living agency we define the kind of organization that is capable of meeting the required conditions for agency (which is not restricted to living organisms). On this basis, we define agency as an autonomous organization that adaptively regulates its coupling with its environment and contributes to sustaining itself as a consequence. We find that spatiality and temporality are the two fundamental domains in which agency spans at different scales. We conclude by giving an outlook for the road that lies ahead in the pursuit of understanding, modeling, and synthesizing agents.

Integrating Autopoiesis and Behavior: An Exploration in Computational Chemo-ethology

Egbert, M. D., Di Paolo, E. — Wed, 23 Sep 2009 02:33:53 PDT

It has been argued that the difference between an autonomous entity and an agent is in the ability of the latter to perform behaviors supplemental to processes of self-maintenance (autopoiesis). Theories have been proposed concerning how such behaviors might relate to autopoiesis, but so far, computational models of autopoiesis have paid little attention to these relations. In this article we present a new model designed to explore the relationship between mechanisms of autopoiesis and behavior. We report on three clarifications of the theory provided by the model: (a) mechanisms of behavior can be related to mechanisms of autopoiesis while remaining operationally distinct, (b) the organization of an operationally closed system can change over time while remaining operationally closed, and (c) behavior modulation based upon autopoietic efficacy has limitations that can be avoided through the use of a partially decoupled behavioral system. Finally, we discuss questions that have surfaced during examination of the model.

The Illusion of Agency: Two Engineering Approaches to Compromise Autonomy and Reactivity in an Artificial System

Aucouturier, J.-J., Ikegami, T. — Wed, 23 Sep 2009 02:33:53 PDT

This article describes and compares two approaches that can be used to build artificial systems that users tend to recognize as agents, based on a review of two systems previously built by the authors. One system, an interactive musical instrument, is a typical artificial intelligence (AI) implementation, based on the technique of constraint programming. The other, a dancing robot, is a typical artificial life (AL) endeavor, specified as a non-linear dynamical system. Although very different in their design, we found that both systems have the same goal of compromising autonomy and reactivity in their user interaction. They elicit interaction and—we argue here—an elusive feeling of agency , because they are neither too predictable nor too random. Creating and controlling such a compromise in a programmatic way is not a trivial problem: we find that both approaches (AI and AL) raise similar pragmatic problems that are in fact rooted in human perception science. Psychological experimentation is needed to clarify the relation between the internal dynamics of the artificial systems and the ongoing feeling of agency (or absence thereof) imparted in their human user. As a first step toward such experimentation, we derive a minimal mathematical model which subsumes both implementations and abstracts them from their respective contexts of music and dance. This model is similar to a Van der Pol oscillator, forced by an input signal coupled to its output in a non-programmatic way. It isolates two critical variables controlling the illusion of agency: the model’s sampling rate and the polynomial order of its reactive term.

Autonomy of Self at Criticality: The Perspective from Synthetic Neuro-Robotics

Tani, J. — Wed, 23 Sep 2009 02:33:53 PDT

This article investigates the phenomenological aspects of selves in relation to autonomous agents. Through a review of a series of neuro-robotics experiments conducted by the author’s group, we elucidate three different aspects of selves, namely, minimal selves, social selves and self-referential selves. Upon integrative discussions of these selves, it is suggested that genuine constructs of "authentic" selves may appear with criticality, which is self-organized in the iterative interplay between regression of past experience and lookahead prediction of future outcomes. It is concluded that genuine autonomy of agents likely originates from genuine autonomy of authentic selves.

On the Role of Social Interaction in Individual Agency

De Jaegher, H., Froese, T. — Wed, 23 Sep 2009 02:33:53 PDT

Is an individual agent constitutive of or constituted by its social interactions? This question is typically not asked in the cognitive sciences, so strong is the consensus that only individual agents have constitutive efficacy. In this article we challenge this methodological solipsism and argue that interindividual relations and social context do not simply arise from the behavior of individual agents, but themselves enable and shape the individual agents on which they depend. For this, we define the notion of autonomy as both a characteristic of individual agents and of social interaction processes. We then propose a number of ways in which interactional autonomy can influence individuals. Then we discuss recent work in modeling on the one hand and psychological investigations on the other that support and illustrate this claim. Finally, we discuss some implications for research on social and individual agency.

Editorial

Di Paolo, E. — Tue, 28 Jul 2009 07:48:02 PDT

Animals Versus Animats: Or Why Not Model the Real Iguana?

Webb, B. — Tue, 28 Jul 2009 07:48:02 PDT

The overlapping fields of adaptive behavior and artificial life are often described as novel approaches to biology. They focus attention on bottom-up explanations and how lifelike phenomena can result from relatively simple systems interacting dynamically with their environments. They are also characterized by the use of synthetic methodologies, that is, building artificial systems as a means of exploring these ideas. Two differing approaches can be distinguished: building models of specific animal systems and assessing them within complete behavior—environment loops; and exploring the behavior of invented artificial animals, often called animats, under similar conditions. An obvious question about the latter approach is, how can we learn about real biology from simulation of non-existent animals? In this article I will argue, first, that animat research, to the extent that it is relevant to biology, should also be considered as model building. Animat simulations do, implicitly, represent hypotheses about, and should be evaluated by comparison to, animals. Casting this research in terms of invented agents serves only to limit the ability to draw useful conclusions from it by deflecting or deferring any serious comparisons of the model mechanisms and results with real biological systems. Claims that animat models are meant to be existence proofs, idealizations, or represent general problems in biology do not make these models qualitatively different from more conventional models of specific animals, nor undermine the ultimate requirement to justify this work by making concrete comparisons with empirical data. It is thus suggested that we will learn more by choosing real, and not made-up, targets for our models.

Animats in the Modeling Ecosystem

Barandiaran, X. E., Chemero, A. — Tue, 28 Jul 2009 07:48:02 PDT

Some Virtues of Modeling With Both Hands

Bechtel, W. — Tue, 28 Jul 2009 07:48:02 PDT

Animals and Animats: Why Not Both Iguanas?

Beer, R. D., Williams, P. L. — Tue, 28 Jul 2009 07:48:02 PDT

In Defense of the Abstracted Animat

Bullock, S. — Tue, 28 Jul 2009 07:48:02 PDT

On Biological Relevance

Etxeberria, A., Moreno, A. — Tue, 28 Jul 2009 07:48:02 PDT

Model the Real, Artificial, or Stylized Iguana? Artificial Life and Adaptive Behavior Can Be Linked Through Pattern-Oriented Modeling

Grimm, V., Railsback, S. F. — Tue, 28 Jul 2009 07:48:02 PDT

Tool-Makers Versus Tool-Users: Division of Labor

Harvey, I. — Tue, 28 Jul 2009 07:48:02 PDT

Let Animats Live!

Hoffmann, M., Pfeifer, R. — Tue, 28 Jul 2009 07:48:02 PDT

Never Mind the Iguana, What About the Tortoise? Models in Adaptive Behavior

Husbands, P. — Tue, 28 Jul 2009 07:48:02 PDT

Rehabilitating Biology as a Natural History

Ikegami, T. — Tue, 28 Jul 2009 07:48:02 PDT

More Synthetic Work Is Needed

Lind, J., Enquist, M. — Tue, 28 Jul 2009 07:48:02 PDT

Iguana Modeling Is Not the Only Game in Town

Noble, J., de Pinedo, M. — Tue, 28 Jul 2009 07:48:02 PDT

No Need for Intellectual Straightjackets

Rohde, M. — Tue, 28 Jul 2009 07:48:02 PDT

Don't Throw the Baby Iguana Out With the Bathwater

Seth, A. — Tue, 28 Jul 2009 07:48:02 PDT

Do Animat Models Always Need a Biological Target Organism?

Wischmann, S. — Tue, 28 Jul 2009 07:48:02 PDT

Response: The Power of Selection

Webb, B. — Tue, 28 Jul 2009 07:48:02 PDT

A Hierarchical Autonomous Robot Controller for Learning and Memory: Adaptation in a Dynamic Environment

Alnajjar, F., Bin Mohd Zin, I., Murase, K. — Tue, 09 Jun 2009 03:23:21 PDT

This work concerns practical issues surrounding the application of learning and memory in a real mobile robot with the goal of optimal navigation in dynamic environments. A novel hierarchical adaptive controller that contains two-level units was developed and trained in a physical mobile robot "e-Puck." In the low-level unit, the robot holds a number of biologically inspired Aplysia -like spiking neural networks that have the property of spike time-dependent plasticity. Each of these networks is trained to become an expert in a particular local environment(s). All the trained networks are stored in a tree-type memory structure that is located in the high-level unit. These stored networks are used as experiences for the robot to enhance its navigation ability in both new and previously trained environments. The robot's memory is designed to hold memories of various lengths and has a simple searching mechanism. Forgetting and dynamic clustering techniques are used to control the memory size. Experimental results show that the proposed model can produce a robot with learning and memorizing capabilities that enable it to survive in complex and highly dynamic environments.

A Steering Taxis Model and the Qualitative Analysis of its Trajectories

Rano, I. — Tue, 09 Jun 2009 03:23:21 PDT

Taxis is a well-known steering technique used by simple animals to approach a stimulus in the real world. However, no mathematical motion model of taxis can be found in the literature. This article derives the differential ordinary equations describing the motion of a Braitenberg vehicle, a widely used taxis model. A qualitative technique for non-linear dynamical systems analysis is applied to investigate the motion generated by the model. Validation of the analysis is performed through several simulations, and conditions for the stimulus source to be reached are obtained. This work fills the theoretical hole in formal models of Braitenberg vehicles and thereby provides theoretical support for the many previous experimental uses of those vehicles for steering tasks.

The Iterated Classification Game: A New Model of the Cultural Transmission of Language

Swarup, S., Gasser, L. — Tue, 09 Jun 2009 03:23:21 PDT

The iterated classification game (ICG) combines the classification game with the iterated learning model (ILM) to create a more realistic model of the cultural transmission of language through generations. It includes both learning from parents and learning from peers. Further, it eliminates some of the chief criticisms of the ILM: that it does not study grounded languages, that it does not include peer learning, and that it builds in a bias for compositional languages. We show that, over the span of a few generations, a stable linguistic system emerges that can be acquired very quickly by each generation, is compositional, and helps the agents to solve the classification problem with which they are faced. The ICG also leads to a different interpretation of the language acquisition process. It suggests that the role of parents is to initialize the linguistic system of the child in such a way that subsequent interaction with peers results in rapid convergence to the correct language.

Re-embodiment of Honeybee Aggregation Behavior in an Artificial Micro-Robotic System

Kernbach, S., Thenius, R., Kernbach, O., Schmickl, T. — Tue, 09 Jun 2009 03:23:21 PDT

In this article we describe the re-embodiment of biological aggregation behavior of honeybees in Jasmine micro-robots. The observed insect behavior, in the context of the insect's sensor—actor system, is formalized as behavioral and motion-sensing meta-models. These meta-models are transformed into a sensor—actor system of micro-robots by means of a sensors virtualization technique. This allows us to keep the efficiency and scalability of the bio-inspired approach. We also demonstrate the systematic character of this re-embodiment procedure on collective aggregation in a real robotic swarm.