Search

In this paper, an approach is advocated to use a hybrid approach towards learning behaviour for computer generated entities (CGEs) in a serious gaming setting. Hereby, an agent equipped with cognitive model is used but this agent is enhanced with Machine Learning (ML) capabilities. This facilitates the agent to exhibit human like behaviour but avoid an expert having to define all parameters explicitly. More in particular, the ML approach utilizes co-evolution as a learning paradigm. An evaluation in the domain of one-versus-one air combat shows promising results.

Volgens schattingen van het Rijksinstituut voor Volksgezondheid en Milieu (RIVM) stijgt het aantal mensen in Nederland met astma en COPD de komende 25 jaar sterk, met respectievelijk 28% en 70%. Dit komt vooral door de bevolkingsgroei en de vergrijzing. Het aantal patiënten met respiratoire allergie (zoals hooikoorts) blijft in deze periode ongeveer gelijk. Deze aandoening komt namelijk bij ouderen minder voor. De verwachting is dat de totale medische kosten voor alle drie de aandoeningen over 25 jaar (fors) zijn gestegen. Voor respiratoire allergie zal dat met 73% zijn; voor astma stijgen de kosten met 150%, voor COPD met 220%. Deze percentages zijn inclusief de jaarlijkse stijging van zorguitgaven door onder andere technologische veranderingen (zoals nieuwe medicijnen) en prijsstijgingen, volgens de trendanalyse van het Centraal Planbureau. Het RIVM heeft deze schattingen gemaakt op verzoek van het Longfonds. De cijfers zijn gebaseerd op nieuwe analyses van de kosten in 2007. Voor astma bedroegen de medische kosten in totaal 287 miljoen euro, gemiddeld 530 euro per patiënt. Dit bedrag bestaat voor bijna driekwart uit kosten voor medicijnen. Bij werknemers komt daar nog gemiddeld 1200 euro per persoon per jaar bovenop vanwege extra ziekteverzuim door astma. Van hen verzuimen werknemers die ouder zijn dan 55 jaar het meest. De medische kosten voor COPD in Nederland in 2007 waren 415 miljoen euro, gemiddeld 1400 euro per patiënt. Hierbij waren geneesmiddelen, ziekenhuisopnames en langdurige zorg (zoals thuiszorg en in verzorgingshuizen) de belangrijke kostenposten. Kosten van arbeidsongeschiktheid waren voor werkenden met COPD gemiddeld 1200 euro per persoon. Voor ziekteverzuim waren deze gemiddeld 1900 euro per werkende met COPD. Deze kosten overtreffen veruit de kosten van het zorggebruik voor COPD. De medische kosten voor respiratoire allergie waren 102 miljoen euro, gemiddeld 170 euro per patiënt. Medicatiekosten vormden hierin het grootste deel, 90%. Er waren te weinig data om de ziekteverzuimkosten betrouwbaar te schatten. De gepresenteerde cijfers over de verwachte stijging van het aantal mensen met deze drie aandoeningen en de kosten die hiermee gemoeid zijn, leveren belangrijke informatie op voor het beleid. Preventie en behandeling zijn daarbij onverminderd belangrijk, zoals stoppen met roken en doelmatiger gebruik van geneesmiddelen. Aangezien er steeds meer oudere patiënten met astma en COPD komen, is specifieke ondersteuning ook voor hen van belang.

Doel van dit project is het bewerkstelligen van verminderen van milieubelasting door effectiever gebruik van middelen en methoden en kennishiaten opheffen rondom gewasbescherming middels netwerken en onafhankelijke kennisdeling (digitaal handboek geïntegreerde gewasbescherming Roos). Dit moet mede leiden tot gedragsverandering bij telers.

In this paper we present an unsupervised learning approach to detect meaningful job traffic patterns in Grid log data. Manual anomaly detection on modern Grid environments is troublesome given their increasing complexity, the distributed, dynamic topology of the network and heterogeneity of the jobs being executed. The ability to automatically detect meaningful events with little or no human intervention is therefore desirable. We evaluate our method on a set of log data collected on the Grid. Since we lack a priori knowledge of patterns that can be detected and no labelled data is available, an unsupervised learning method is followed. We cluster jobs executed on the Grid using Affinity Propagation. We try to explain discovered clusters using representative features and we label them with the help of domain experts. Finally, as a further validation step, we construct a classifier for five of the detected clusters and we use it to predict the termination status of unseen jobs.

In this chapter an agent-based ambient system is presented to support persons in learning specific movement patterns. The ambient system serves as a personal coach that observes a person’s movement pattern, and analyses this based on comparison with an ideal pattern generated by optimisation using a computational musculoskeletal model for this type of pattern minimizing knee joint loading. Based on this analysis the Personal Coach generates advice to adapt the person’s pattern in order to better approximate the ideal pattern. The Personal Coach has been designed using the agent design method DESIRE, thereby reusing an available generic agent model. The system was evaluated (a proof of principle) by setting up an environment in which sensoring of body part positions was incorporated. In evaluations with a few subjects substantial improvement of the movement pattern compared to the ideal movement pattern was achieved.

When considering intelligent agents that interact with humans, having an idea of the trust levels of the human, for example in other agents or services, can be of great importance. Most models of human trust that exist, are based on some rationality assumption, and biased behavior is not represented, whereas a vast literature in Cognitive and Social Sciences indicates that humans often exhibit non-rational, biased behavior with respect to trust. This paper reports how some variations of biased human trust models have been designed, analyzed and validated against empirical data. The results show that such biased trust models are able to predict human trust significantly better. © 2011 IEEE.

Cyclophosphamide is an alkylating agent used in antineoplastic and immunosuppressive therapies. Symptomatic hyponatraemia is a rare but life-threatening complication in patients treated with cyclophosphamide. We report the case of a 64-year-old woman with breast cancer who developed severe symptomatic hyponatraemia with a generalised seizure and convulsions after a second cycle of adjuvant chemotherapy with 5-fluouracil, epirubicin and cyclophosphamide. She completely recovered after correction of the serum sodium concentration without neurological deficits. Physicians prescribing cyclophosphamide, irrespective of the treatment indication and dosage, should be aware of this potentially life-threatening complication.

In this paper, an integrated agent model is introduced addressing mutually interacting Situation Awareness and Functional State dynamics in decision making. This shows how a human's functional state, more specific a human's exhaustion and power, can influence a human's situation awareness, and in turn the decision making. The model is illustrated by a number of simulation scenarios. © 2011 Springer-Verlag.

In order for agents to be able to act intelligently in an environment, a first necessary step is to become aware of the current situation in the environment. Forming such awareness is not a trivial matter. Appropriate observations should be selected by the agent, and the observation results should be interpreted and combined into one coherent picture. Humans use dedicated mental models which represent the relationships between various observations and the formation of beliefs about the environment, which then again direct the further observations to be performed. In this paper, a generic agent model for situation awareness is proposed that is able to take a mental model as input, and utilize this model to create a picture of the current situation. In order to show the suitability of the approach, it has been applied within the domain of F-16 fighter pilot training for which a dedicated mental model has been specified, and simulations experiments have been conducted.

Trust dynamics is often modelled in relation to experiences and usually from a cognitive perspective, leaving affective aspects out of consideration. Inspired by neurological findings that show more and more how in mental processes cognitive and affective aspects are intertwined, in this paper, a model for trust dynamics is introduced incorporating the relation between trust and feeling. The model makes use of a Hebbian learning principle and describes how trust depends on experiences viewed as information obtained over time, but also as emotional responses and feelings.

In this paper, the results of a validation experiment for two existing computational trust models describing human trust are reported. One model uses experiences of performance in order to estimate the trust in different trustees. The second model in addition carries the notion of relative trust. The idea of relative trust is that trust in a certain trustee not solely depends on the experiences with that trustee, but also on trustees that are considered competitors of that trustee. In order to validate the models, parameter adaptation has been used to tailor the models towards human behavior. A comparison between the two models has also been made to see whether the notion of relative trust describes human trust behavior in a more accurate way. The results show that taking trust relativity into account indeed leads to a higher accuracy of the trust model. Finally, a number of assumptions underlying the two models are verified using an automated verification tool. © 2011 International Federation for Information Processing.

In this chapter, we propose to outline the scientific area that addresses Ambient Intelligence applications in which not only sensor data, but also knowledge from the human-directed sciences such as biomedical science, neuroscience, and psychological and social sciences is incorporated. This knowledge enables the environment to perform more in-depth, human-like analyses of the functioning of the observed humans, and to come up with better informed actions. A structured approach to embed human knowledge in Ambient Intelligence applications is presented an illustrated using two examples, one on automated visual attention manipulation, and another on the assessment of the behaviour of a car driver.

Models of agents or multiagent systems in a certain application area can be made at different levels of abstraction. The aim of this paper is to clarify different dimensions of abstraction. The three dimensions of considered are: the process abstraction dimension, the temporal dimension, and the agent cluster dimension. Thus a three-dimensional framework is obtained in which different types of multi-agent system models can be classified. For a number of multiagent system models in different application areas from the literature it is discussed how they fit in the framework.

Behavioural activation treatment has been found to be an effective psychological treatment for depression, also if delivered as self-administered psychotherapy via the internet. However, the role of supporting professionals remains important for successful application of the therapy. In this paper a system is presented that delivers automated behavioural activation therapy via both a mobile phone and a personal website. The system motivates the client to continue with the treatment and helps him/her through the different procedures of the treatment. The architecture of the system follows a generic ambient agent architecture. A first pilot study of the system indicates that it is technically feasible and perceived as useful.

Within cognitive models, desires are often considered as functional concepts that play a role in efficient focusing of behaviour. In practice a desire often goes hand in hand with having certain feelings. In this paper by adopting neurological theories a model is introduced incorporating both cognitive and affective aspects in the dynamics of desiring and feeling. Example simulations are presented, and both a mathematical and logical analysis is included.