Evolution of a predictive internal model in an embodied and situated agent

We show how simulated robots evolved for the ability to display a context-dependent periodic behavior can spontaneously develop an internal model and rely on it to fulfill their task when sensory stimulation is temporarily unavailable. The analysis of some of the best evolved agents indicates that their internal model operates by anticipating sensory stimuli. More precisely, it anticipates functional properties of the next sensory state rather than the exact state that sensors will assume. The characteristics of the states that are anticipated and of the sensorimotor rules that determine how the agents react to the experienced states, however, ensure that they produce very similar behaviour during normal and blind phases in which sensory stimulation is available or is self-generated by the agent, respectively. Agents’ internal models also ensure an effective transition during the phases in which agents’ internal dynamics is decoupled and re-coupled with the sensorimotor flow. Our results suggest that internal models might have arisen for behavioral reasons and successively exapted for other cognitive functions. Moreover, the obtained results suggest that self-generated internal states should not necessarily match in detail the corresponding sensory states and might rather encode more abstract and motor-oriented information.     

Emergence of rules in cell society: Differentiation, hierarchy, and stability

A dynamic model for cell differentiation, where cells with internal chemical reaction dynamics interact with each other and replicate was studied. It led to spontaneous differentiation of cells and determination, as discussed in the isologous diversification. The following features of the differentiation were obtained: (1) hierarchical differentiation from a ’stem’ cell to other cell types, with the emergence of the interaction-dependent rules for differentiation; (2) global stability of an ensemble of cells consisting of several cell types, that were sustained by the emergent, autonomous control on the rate of differentiation; (3) existence of several cell colonies with different cell-type distributions. The results provide a novel viewpoint on the origin of a complex cell society, while relevance to some biological problems, especially to the hemopoietic system, is also discussed.     

From sphere to torus: A topological view of the metazoan body plan

From the viewpoint of mathematical topology, membrane systems in intact living cells can be described as closed and orientable surfaces, i.e., as surfaces with two sides and no boundary lines so that an ‘inside’ and an ‘outside’ can be distinguished. Usually, biomembranes represent topological spheres, often one embedded in another one. Toroidal membranes are occasionally observed, e.g., in specialized structures of plant cells like the prolamellar body. Here, we propose that rules analogous to those that govern the topology of biomembranes hold for the epithelial cell sheets that cover anatomically external as well as internal surfaces of multicellular animals. We suggest to study the emergence of morphological complexity during metazoan development using concepts from mathematical topology, and propose experimental analyses of those topological transitions that appear to be relevant in development and evolution. Based on a lecture held by H J at the Max Planck-Institute for Physics of Complex Systems, Dresden, 26 April 2001.     

Risk-Pooling and Herd Survival: An Agent-Based Model of a Maasai Gift-Giving System

We use agent-based modeling to study osotua, a gift giving system used by the Maasai of East Africa. Osotua’s literal meaning is “umbilical cord,” but it is used metaphorically to refer to a specific type of gift-giving relationship. Osotua relationships are characterized by respect, responsibility and restraint. Osotua partners ask each other for help only if they are in need and provide help only when asked and only if they are able. We hypothesize that under the ecologically volatile conditions in which Maasai pastoralists have traditionally lived, such a system is particularly suited to risk pooling. Here we explore whether osotua increases the viability of herds by comparing herd survivorship and stability under osotua rules to a) no exchange and b) probabilistic rules for requesting and giving livestock. Results from this model suggest that this gift-giving system can dramatically increase herd longevity through a limited pooling of risk.     

Spike timing dependent synaptic plasticity in biological systems

 Association of a presynaptic spike with a postsynaptic spike can lead to changes in synaptic efficacy that are highly dependent on the relative timing of the pre- and postsynaptic spikes. Different synapses show varying forms of such spike-timing dependent learning rules. This review describes these different rules, the cellular mechanisms that may be responsible for them, and the computational consequences of these rules for information processing and storage in the nervous system. Received: 16 January 2002 / Accepted: 3 June 2002 Acknowledgements. This research is supported in part by a National Science Foundation grant IBN 98-08887 (awarded to PDR), and by National Institutes of Health grants R01-MH49792 (awarded to CCB), R01-MH60996 (awarded to CCB), and R01-MH60996 (awarded to PDR). Correspondence to: P. D. Roberts (e-mail: robertpa@ohsu.edu)     

Applying voting rules to panel-based decision making in LCA

Background, aim, and scope  Cross-category weighting is one possible way to facilitate internal decision making when dealing with ambiguous impact assessment results, with simple additive weighting being a commonly used method. Yet, the question as to whether the methods applied today can, in fact, identify the most “environmentally friendly” alternative from a group perspective remains unanswered. The aim of this paper is to propose a new method for group decision making that ensures the effective identification of the most preferable alternative. Materials and methods  Common approaches to deduce a single set of weighting factors for application in a group decision situation (e.g., arithmetic mean, consensus) are discussed based on simple mathematics, empirical data, and thought experiments. After proposing an extended definition for “effectiveness” in group decision making, the paper recommends the use of social choice theory whose main focus is to identify the most preferable alternative based on individuals’ rankings of alternatives. The procedure is further supplemented by a Monte Carlo analysis to facilitate the assessment of the result’s robustness. Results  The general feasibility of the method is demonstrated. It generates a complete ranking of alternatives, which does not contain cardinal single scores. In terms of effectiveness, the mathematical structure of the procedure ensures the eligibility for compromise of the group decision proposal. The sensitivity analysis supports the decision makers in understanding the robustness of the proposed group ranking. Discussion  The method is based upon an extended definition of effectiveness which acknowledges the eligibility for compromise as the core requirement in group decision contexts. It is shown that multi-attribute decision-making (MADM) methods in use in life cycle assessment (LCA) today do not necessarily meet this requirement because of their mathematical structure. Further research should focus on empirical proof that the generated group results are indeed more eligible for compromise than results generated by current methods that utilize an averaged group weighting set. This is closely related to the question considering under which mathematical constraints it is even possible to generate an essentially different result. Conclusions  The paper describes a new multi-attribute group decision support system (MGDSS) for the identification of the most preferable alternative(s) for use in panel-based LCA studies. The main novelty is that it refrains from deducing a single set of weighting factors which is supposed to represent the panel as a whole. Instead, it applies voting rules that stem from social choice theory. Because of its mathematical structure, the procedure is deemed superior to common approaches in terms of its effectiveness. Recommendations and perspectives  The described method may be recommended for use in internal, panel-based LCA studies. In addition, the basic approach of the method—the combination of MADM methods with social choice theory—can be recommended for use in all those situations where multi-attribute decisions are to be made in a group context.     

Mathematical models of eye movements in reading: a possible role for autonomous saccades

 An efficient method for the exact numerical simulation of semi-Markov processes is used to study minimal models of the control of eye movements in reading. When we read a text, typical sequences of fixations form a rather complicated trajectory – almost like a random walk. Mathematical models of eye movement control can account for this behavior using stochastic transition rules between few discrete internal states, which represent combinations of certain stages of lexical access and saccade programs. We show that experimentally observed fixation durations can be explained by residence-time-dependent transition probabilities. Stochastic processes with this property are known as semi-Markov processes. For our numerical simulations we use the minimal process method (Gillespie algorithm), which is an exact and efficient simulation algorithm for this class of stochastic processes. Within this mathematical framework, we study different forms of coupling between eye movements and shifts of covert attention in reading. Our model lends support to the existence of autonomous saccades, i.e., the hypothesis that initiations of saccades are not completely determined by lexical access processes. Received: 21 March 2000 / Accepted in revised form: 10 January 2001     

Connectivity and information transfer in flow networks: Two magic numbers in ecology?

An ecosystem can be visualized as a graph of certain preassigned trophic compartments; these nodes are then mutually connected through the internal exchanges of material and energy. The mathematical theory of information can be applied to such a graph in order to define two relevant indices: a measure of connectivity (the entropy H of the connections) and a measure of the degree of the “energetic” specialization (the internal transfer of informationI). The computation of these indices in stationary real cases suggests that the observed complexity of ecosystems is conditioned by two competing effects. The first can be interpreted as a “thermodynamical” principle related to the unavoidable irreversibility taking place inside the system, whereas the second can be taken as a “biological” principle concerned with the selection of some particular interactions: those which maximize the information circulating between the compartments.     

Hierarchical graphs for rule-based modeling of biochemical systems

Background  

In rule-based modeling, graphs are used to represent molecules: a colored vertex represents a component of a molecule, a vertex attribute represents the internal state of a component, and an edge represents a bond between components. Components of a molecule share the same color. Furthermore, graph-rewriting rules are used to represent molecular interactions. A rule that specifies addition (removal) of an edge represents a class of association (dissociation) reactions, and a rule that specifies a change of a vertex attribute represents a class of reactions that affect the internal state of a molecular component. A set of rules comprises an executable model that can be used to determine, through various means, the system-level dynamics of molecular interactions in a biochemical system.     

Expertise in expert systems: knowledge acquisition for biological expert systems

In this paper it is argued that an expert system requires morethan factual knowledge before it can display expertise in agiven domain. The additional knowledge consists of the heuristicsor ‘rules of thumb’ used by an expert to manipulateand interpret the factual knowledge. The knowledge acquisitionphase of an expert system project involves determining the factualknowledge (which may be obtained from published sources) andthe heuristics used by an expert to manipulate that knowledge-theseheuristics can only be obtained from an expert. In reviewingexisting biological expert systems it is apparent that manycontain only the factual knowledge relating to the domain, andlack the heuristics that enable such systems to show expertise.This paper reviews a number of knowledge acquisition techniqueswhich could be used for acquiring heuristic knowledge and discusseswhen their use is appropriate. The knowledge acquisition techniquesdiscussed are those suitable for the development of small-scaleexpert systems as these are most likely to be of interest tobiologists. The techniques include the use of questionnaires,interview techniques and protocol analysis; particular emphasisis placed on a mod cation to the ‘twenty questions’interview technique which was developed specifically to elicittaxonomic knowledge relating to water mite identification.     

The PhyloCode is fatally flawed, and the “Linnaean” System can easily be fixed

Promoters of the PhyloCode have mounted an intensive and deceptive publicity campaign. At the centerpiece of this campaign have been slogans such as that the Linnaean System will “goof you up,” that the PhyloCode is the “greatest thing since sliced bread,” and that systematists are “afraid” to propose new names because of “downstream consequences.” Aside from such subscientific spin and sloganeering, proponents of the PhyloCode have offered nothing real to back up claims of greater stability for their new system. They have also misled many into believing that the PhyloCode is the only truly phylogenetic system. The confusion that has been fostered involves several discrete arguments, concerning: a new “method” of “designating” names, rank-free taxonomy, uninomial nomenclature, and issues of priority. Claims that the PhyloCode produces a more stable nomenclature are false, as shown with the example of “paleoherbs.” A rank-free system of naming requires an annotated reference tree for even the simplest exchanges of information. This would be confusing at best and would cripple our ability to teach, learn, and use taxonomic names in the field or in publications. We would be confronted by a mass of polynomial names, tied together only by a tree graphic, with no agreed name (except a uninomial, conveying no hierarchy) to use for any particular species. The separate issue of stability in reference to rules of priority and rank can be easily addressed within the current codes, by implementation of some simple changes, as we will propose in this article. Thus there is no need to “scrap” the current Linnaean codes for a poorly reasoned, logically inconsistent, and fatally flawed new code that will only bring chaos.     

Ranking the importance of alerts for problem determination in large computer systems

The complexity of large computer systems has raised unprecedented challenges for system management. In practice, operators often collect large volume of monitoring data from system components and set up many rules to check data and trigger alerts. However, the alerts from various rules usually have different problem reporting accuracy because their thresholds are often manually set based on operators’ experience and intuition. Meantime, due to system dependencies, a single problem may trigger many alerts at the same time in large systems and the critical question is which alert should be analyzed first in the following problem determination process. In this paper, we propose a novel peer review mechanism to rank the importance of alerts and the top ranked alerts are more likely to be true positives. After comparing a metric value against its threshold to generate alerts, we also compare the value with the equivalent thresholds from many other rules to determine the importance of alerts. Our approach is evaluated with a real test bed system and experimental results are also included to demonstrate its effectiveness.     

Control model of human stance using fuzzy logic

 A control model of human stance is proposed based on knowledge from behavioral experiments and physiological systems. The proposed model is based on the control of global variables specific to body orientation and alignment, rather than on the control of the body’s center of mass within the base of support. Furthermore, the proposed control model is not based on purely inverted pendulum body mechanics where only motion at one joint is controlled, as for instance the ankle. In the proposed model, the degrees of freedom are controlled by using reciprocal and synergistic muscle actions at multiple joints. The control model is based on three sets of different global variables which act in parallel: (1) limb length and its derivative, (2) limb orientation and its derivative, and (3) trunk attitude and its derivative. An important feature of the control model is the use of fuzzy logic, which enables us to model experimental findings and physiological knowledge in a meaningful and explicit way using fuzzy if-then rules. In the control model, 36 fuzzy if-then rules are implemented and applied using a four-linked segment model consisting of a trunk, thigh, shank and foot. Uni- and biarticular limb muscles and trunk muscles are represented as torque actuators at each individual joint. In the model, three sets of global variables act in parallel and make corrective and coordinated responses to internal, self-induced perturbations. The data show that the use of global variables and fuzzy logic successfully enables us to model human standing with sway about a point of equilibrium. Small changes in, for example, total body sway are comparable to those seen during natural sway in human stance. The selected controllers—limb length, limb orientation and trunk attitude—seem to be appropriate for human stance control. Received: 30 October 1996/Accepted in revised form: 7 April 1997     

Convectively stable pressure profile in magnetic confinement systems with internal rings

A convectively stable pressure profile in a long multiple-mirror (corrugated) magnetic confinement system with internal current-carrying rings is calculated. The plasma energy content in the axial region can be increased by using an internal ring that reverses the on-axis magnetic field direction and gives rise to an average magnetic well near the axis. The pressure profile in the outer region—outside the magnetic well—is considered in detail. It is shown that, in the radial pressure profile, a pedestal can be formed that leads to a higher pressure drop between the center and the plasma edge. The pressure profile is calculated from the Kruskal-Oberman criterion—a necessary and sufficient condition for the convective stability of a collisionless plasma. The revealed pedestal arises near the boundary of the average magnetic well in the region of the smallest but alternating-sign curvature of the magnetic field lines due to a break in the convectively stable pressure profile. Such a shape of the stable pressure profile can be attributed to the stabilizing effect of the alternating-sign curvature of the field lines in the multiple-mirror magnetic confinement systems under consideration.     

Criminal Law as It Pertains to Patients Suffering from Psychiatric Diseases

The McNaughton rules for determining whether a person can be successfully defended on the grounds of mental incompetence were determined by a committee of the House of Lords in 1843. They arose as a consequence of the trial of Daniel McNaughton for the killing of Prime Minister Sir Robert Peel’s secretary. In retrospect it is clear that McNaughton suffered from schizophrenia. The successful defence of McNaughton on the grounds of mental incompetence by his advocate Sir Alexander Cockburn involved a profound shift in the criteria for such a defence, and was largely based on the then recently published “scientific” thesis of the great US psychiatrist Isaac Ray, entitled A Treatise on the Medical Jurisprudence of Insanity. Subsequent discussion of this defence in the House of Lords led to the McNaughton rules, still the basis of the defence of mental incompetence in the courts of much of the English-speaking world. This essay considers one of these rules in the light of the discoveries of cognitive neuroscience made during the 160 years since Ray’s treatise. A major consideration is the relationship between “the power of self-control” and “irresistible impulse” as conceived by Cockburn on the one hand, and by cognitive neuroscience on the other. The essay concludes with an analysis of the notion of “free will” and of the extent to which a subject can exert restraint in the absence of particular synaptic connections in the brain.     

Quality leadership and quality control

Different quality control rules detect different analytical errors with varying levels of efficiency depending on the type of error present, its prevalence and the number of observations. The efficiency of a rule can be gauged by inspection of a power function graph. Control rules are only part of a process and not an end in itself; just as important are the trouble-shooting systems employed when a failure occurs. 'Average of patient normals' may develop as a usual adjunct to conventional quality control serum based programmes. Acceptable error can be based on various criteria; biological variation is probably the most sensible. Once determined, acceptable error can be used as limits in quality control rule systems.

A key aspect of an organisation is leadership, which links the various components of the quality system. Leadership is difficult to characterise but its key aspects include trust, setting an example, developing staff and critically setting the vision for the organisation. Organisations also have internal characteristics such as the degree of formalisation, centralisation, and complexity. Medical organisations can have internal tensions because of the dichotomy between the bureaucratic and the shadow medical structures.

     

Minimising inter-laboratory variation when constructing a unified molecular database of plant varieties in an allogamous crop

The construction of large-scale databases of molecular profiles of plant varieties for variety identification and diversity analyses is of considerable interest. When varieties of an allogamous species such as oilseed rape are analysed and described using molecular markers such as microsatellites, care is needed to represent the variety in a meaningful yet useful way. It is possible to characterise such heterogeneous genotypes by analysing bulked samples comprising more than one individual seed or plant, but this approach may result in complex microsatellite profiles. Intuitively it would be reasonable to represent a variety by the common ‘major alleles’ in a profile, but how to define these ‘major alleles’ remains problematic. This paper describes methods of analysing DNA microsatellite data that will allow independent and objective data production at a number of laboratories. Methods for establishing allele scoring rules (thresholding) are described and the effect of these rules on the utility of the data is discussed.     

Role of Utility and Inference in the Evolution of Functional Information

Functional information means an encoded network of functions in living organisms from molecular signaling pathways to an organism’s behavior. It is represented by two components: code and an interpretation system, which together form a self-sustaining semantic closure. Semantic closure allows some freedom between components because small variations of the code are still interpretable. The interpretation system consists of inference rules that control the correspondence between the code and the function (phenotype) and determines the shape of the fitness landscape. The utility factor operates at multiple time scales: short-term selection drives evolution towards higher survival and reproduction rate within a given fitness landscape, and long-term selection favors those fitness landscapes that support adaptability and lead to evolutionary expansion of certain lineages. Inference rules make short-term selection possible by shaping the fitness landscape and defining possible directions of evolution, but they are under control of the long-term selection of lineages. Communication normally occurs within a set of agents with compatible interpretation systems, which I call communication system. Functional information cannot be directly transferred between communication systems with incompatible inference rules. Each biological species is a genetic communication system that carries unique functional information together with inference rules that determine evolutionary directions and constraints. This view of the relation between utility and inference can resolve the conflict between realism/positivism and pragmatism. Realism overemphasizes the role of inference in evolution of human knowledge because it assumes that logic is embedded in reality. Pragmatism substitutes usefulness for truth and therefore ignores the advantage of inference. The proposed concept of evolutionary pragmatism rejects the idea that logic is embedded in reality; instead, inference rules are constructed within each communication system to represent reality, and they evolve towards higher adaptability on a long time scale.     

Modeling the hypothalamus–pituitary–adrenal system: homeostasis by interacting positive and negative feedback

The hypothalamus–pituitary–adrenal (HPA) system is closely related to stress and the restoration of homeostasis. This system is stimulated in the second half of the night, decreases its activity in the daytime, and reaches the homeostatic level during the late evening. In this paper, we derive and discuss a novel model for the HPA system. It is based on three simple rules that constitute a principle of homeostasis and include only the most substantive physiological elements. In contrast to other models, its main components include, apart from the conventional negative feedback ingredient, a positive feedback loop. To validate the model, we present a parameter estimation procedure that enables one to adapt the model to clinical observations. Using this methodology, we are able to show that the novel model is capable of simulating clinical trials. Furthermore, the stationary state of the system is investigated. We show that, under mild conditions, the system always has a well-defined set-point, which reflects the clinical situation to be modeled. Finally, the computed parameters may be interpreted from a physiological point of view, thereby leading to insights about diseases like depression, obesity, or diabetes.