Abstract: Agents in an organization need to coordinate their actions in order to
reach the organizational goals. This research describes the relation between types
of coordination and the autonomy of actors. In an experimental setting we show
that there is not one best way to coordinate in all situations. The dynamics and
complexity of, for example, crisis situations require a crisis management orga-
nization to work with dynamic types of coordination. In order to reach dynamic
coordination we provide the actors with adjustable autonomy. Actors should be
able to make decisions at different levels of autonomy and reason about the re-
quired level. We propose a way to implement this in a multi-agent system. The
agent is provided with reasoning rules with which it can control the external in-
ﬂuences on its decision-making.
Abstract: Agents in an organization need to coordinate their actions in order to reach the organizational goals. This research describes the relation between types of coordination and the autonomy of actors. In an experimental setting we show that there is not one best way to coordinate in all situations. The dynamics and complexity of, for example, crisis situations require a crisis management orga- nization to work with dynamic types of coordination. In order to reach dynamic coordination we provide the actors with adjustable autonomy. Actors should be able to make decisions at different levels of autonomy and reason about the re- quired level. We propose a way to implement this in a multi-agent system. The agent is provided with reasoning rules with which it can control the external in- ﬂuences on its decision-making.
Abstract: The multidimensional assignment problem (MAP) is a combinatorial optimization problem arising in many applications, for instance in multi-target multi-sensor tracking problems. It is well-known that the MAP is NP-hard. The objective of a MAP is to match d-tuples of objects in such a way that the solution with the optimum total cost is found. In this paper a new class of approximation algorithms to solve the MAP is presented, named K-SGTS, and its effectiveness in multi-target multi-sensor tracking situations is shown. Its computational complexity is proven to be polynomial. Experimental results on the accuracy and speed of K-SGTS are provided in the last section of the paper
Abstract: The collaboration between humans (actors) and artificial entities (agents) can be a potential performance boost.
Agents, as complementary artificial intelligent entities, can alleviate actors from certain activities, while enlarging
the collective effectiveness. This paper describes our approach for experimentation with actors, agents and their
interaction. This approach is based on a principled combination of existing empirical research methods and is
illustrated by a small experiment which assesses the performance of a specific actor-agent team in comparison with
an actor-only team in an incident management context. The REsearch and Simulation toolKit (RESK) is
instrumental for controlled and repeatable experimentation. The indicative findings show that the approach is viable
and forms a basis for further data collection and comparative experiments. The approach supports applied actor-
agent research to show its (dis)advantages as compared to actor-only solutions.
Abstract: We introduce a support system concept that offers both work-centered and human-aware support for operators in tactical command and control environments. The support system augments the cognitive capabilities of the operator by offering instant, personalized task and work support. The operator obtains support by entering into a collaborative agreement with support agents. Such an agreement creates opportunities to establish adaptive capabilities and human-aware support features. We describe the concept in and propose an experimental design to evaluate its effectiveness in tactical environments.
Abstract: This paper studies the use of agent communication in ubiquitous computing. This application domain allows us to investigate the efficient handling of large quantities of information in agent-based systems. We will present an approach to dynamically set up a communication network between agents which aims to minimize the communication load. The approach is based on a formal ontological notion of informativeness, on quantitative measures such as information gain and on the proper use of interaction mechanisms such as Publish/Subscribe. We also present experimental results which have been obtained using our prototyping tool called Ubismart.
Abstract: An experimental automated dialogue system that plays the role of a crisis hotline dispatcher is currently developed. Besides controlling the communication flow, this system is able to retrieve information about crisis situations from user's input. It offers a natural user interaction by the ability to perceive and respond to human emotions. The system has an emotion recognizer that is able to recognize the emotional loading from user's linguistic content. The recognizer uses a database that contains selected keywords on a 2D "arousal" and "valence" scale. The output of the system provides not only the information about the user's emotional state but also an indication of the urgency of his/her information regarding to crisis. The dialogue system is able to start a user friendly dialogue, taking care of the content, context and emotional loading of user's utterances.
Abstract: To encounter limited options of user interactions provided by PDAs, we propose a new interaction paradigm using icons to represent concepts or ideas. As icons offer a potential across language barriers, we developed two experimental iconbased communication interfaces that are language independent, as proofs of our concept. They were applied on a tool for travellers. Users can create an iconic message as
realization of their concepts or ideas in mind. The tool is able to interpret and convert the message to (natural language) text and speech. Corpus-based approach was used to have more insight of how humans express their concepts or ideas using this type of messages. This tool also provides an icon
prediction to have faster interaction in next icon selections.
Our user test results confirmed that using provided icons our target users could express their concepts and ideas solely using a spatial arrangement of icons.
Abstract: Our work addresses the problem of autonomous concept formation from a design point of view, providing an initial answer to the question: What are the design features of an architecture supporting the acquisition of different types of concepts by an autonomous agent?
Autonomous agents, that is systems capable of interacting independently with their environment in the pursuit of their own goals, will provide the framework in which we study the problem of autonomous concept formation. Humans and most animals may in this sense also be regarded as autonomous agents, but our concern will be with artiﬁcial autonomous agents. A detailed survey and discussion of the many issues surrounding the notion of ‘artiﬁcial agency’ is beyond the scope of this thesis and a good overview can be found in [Wooldridge and Jennings, 1995]. Instead we will focus on how artiﬁcial agents could be endowed with representational and modelling capabilities.
The ability to form concepts is an important and recognised cognitive ability, thought to play an essential role in related abilities such as categorisation, language understanding, object identiﬁcation and recognition, reasoning, all of which can be seen as different aspects of intelligence. Concepts and categories are studied within cognitive science, where scientists are concerned with human conceptual abilities and mental representations of categories, but they have been addressed also in the rather different domain of machine learning and classiﬁcatory data analysis, where the focus is on the development of algorithms for clustering problems and induction problems [Mechelen et al., 1993]. The two ﬁelds are well distinct and only recently have started to interact, but even though the importance of concepts have been recognised, the nature of concepts is controversial, in the sense that there is no commonly agreed theory of concepts, and it is still far from obvious which representational means are most suited to capture the many cognitive functions that concepts are involved in.
Among the goals of this thesis there is the attempt to bring together different lines of argumentation that have emerged within philosophy, cognitive science and AI, in order to establish a solid foundation for further research into the representation and acquisition of concepts by autonomous agents. Thus, our results and conclusions will often be stated in terms of new insights and ideas, rather than resulting in new algorithms or formal methods.
Our focus will be on affordance concepts — discussed in detail in Chapter 4 — and our main contributions will be:
* An argument showing that concepts should be thought of as belonging to different kinds, where the differences among these kinds are to be captured in terms of architecture features supporting their acquisition.
* A description (and partial implementation) of a minimal architecture (the Innate Adaptive Behaviour architecture – IAB architecture for short) supporting the acquisition of affordance concepts; the IAB architecture is actually a proposal for a sustaining mechanism, in the sense of [Margolis, 1999], for affordances, and makes clear the necessity of a minimal structure for the representation of affordances.
When addressing concept formation in AI, what can be called the ‘system level’ is often overlooked, which means that concepts and categories are rarely studied from the point of view of a system, autonomous and complete, that might need such constructs and can acquire them only by means of interactions with its environment, under the constraints of its cognitive architecture. Also within psychology, the focus is usually on structural aspects of concepts rather than on developmental issues [Smith and Medin, 1981]. Our approach – an architecture-based approach – is an attempt (i) to show that a system level perspective on concept formation is indeed possible and worth exploring, and (ii) to provide an initial, maybe simple, but concrete example of the insights that can be gained from such an approach. Since the methodology that we propose to study concept formation is a general one, and can be applied also to other types of concepts, we decided to mention broadly ‘autonomous concept formation’ rather than ‘autonomous affordance-concepts formation’ in the title of the thesis.
Abstract: This deliverable explores basic characteristics of human groups and teams in order to derive implications for actor-agent teams (AAT’s). From a socio-psychological group dynamics perspective group developmental stages, membership, cohesion, subgroups, social status, roles, norms and leadership are defined and explained in order to enhance the understanding of the processes that are part of human group behavior. The document subsequently briefly explains what ‘actor-agent-team’ means, making the assumption that the factors that play a role in human-only teams also play a role in AAT’s and putting further implications to discussion.
Abstract: Reinforcement learning (RL) is a widely used
paradigm for learning control. Computing exact RL solutions is
generally only possible when process states and control actions
take values in a small discrete set. In practice, approximate
algorithms are necessary. In this paper, we propose an approximate, model-based Q-iteration algorithm that relies on
a fuzzy partition of the state space, and a discretization of
the action space. Using assumptions on the continuity of the
dynamics and of the reward function, we show that the resulting
algorithm is consistent, i.e., that the optimal solution is obtained
asymptotically as the approximation accuracy increases. An
experimental study indicates that a continuous reward function
is also important for a predictable improvement in performance
as the approximation accuracy increases.
Abstract: Decision making during crises takes place in (multi-agency) teams, in a bureaucratic political context. As a
result, the common notion that during crises decision making should be done in line with a Command & Control
structure is invalid. This paper shows that the best way for crisis decision making teams in a bureaucratic
political context is to follow an integrative negotiation approach as the shared mental model of decision making.
This conclusion is based on an analysis of crisis decision making by teams in a bureaucratic political context.
First of all this explains why in a bureaucratic political context the Command & Control adage does not hold.
Secondly, this paper motivates why crisis decision making in such context can be seen as a negotiation process.
Further analysis of the given context shows that an assertive and cooperative approach suits crisis decision
Abstract: In complex strategic decision-making situations the need for well-structured support arises. To evaluate decision alternatives, information about the situation and its development must be determined, managed and processed by the best available experts. For various types of information different reasoning principles have been developed: deterministic, probabilistic, fuzzy and techniques for reasoning under ignorance (i.e., the likelihood of an event cannot be quantified). We propose a new approach based on Decision Maps supporting decision makers under fundamental uncertainty by generating descriptions of different possible situation developments (scenarios) in a distributed manner. The scenarios are evaluated using Multi-Criteria Decision Analysis techniques.
Abstract: We derive a number of well known deterministic latent variable
models such as PCA, ICA, EPCA, NMF and PLSA as variational
EM approximations with point posteriors. We show that the often
practiced heuristic of “folding-in” can lead to overly optimistic
estimates of the test-set log-likelihood and we verify this result
experimentally. We trace this problem back to an infinitely negative
entropy term that is ignored in the variational approximation.
Abstract: Agents that are acting in a multi-agent system can be inﬂuenced by other agents via interactions. At the same time
agents are expected to be autonomous; they have control
over their internal state and their behavior. Our claim is
that autonomous agents should be aware of the external inﬂuences they are sensitive for. This paper describes diﬀerent types of inter-agent inﬂuences and investigates their role in
the reasoning process of an agent. We propose away to
represent the experienced inﬂuences in the mental model of
an agent. By doing so we provide the awareness of external inﬂuences and guarantee the autonomy requiremen of
Abstract: This report describes strengths and limitations of human cognition in relation to computers and automation. In this sense, the report may be interesting in any context in which humans work together with (intelligent) machines. A theoretical standpoint is presented, which allows useful interpretation of experimental results on human cognition. A method is presented by which human cognitive strength can be measured.
Abstract: In recent years, we have developed an icon-based communication interface to represent concepts and ideas. Users can create messages to communicate with others using a spatial arrangement of visual symbols. We deployed our icon-based interface in a serious game environment of a disaster and rescue simulator, which is capable of simulating real disaster situations using information from human user observers' reports. To support faster interaction, we designed a highly adaptive interface for optimizing the next icon look up. Inspired by the Fitaly keyboard, the system rearranges the icons menu's layout dynamically to minimize the searching time. Users are able to find their desired icons fast since the next icon selection is most likely to be one that is (on or) around the center. Our user tests showed that the developed icon-based interface could serve as a communication mediator. The experimental results also indicated that the Fitaly-based interface allowed for much faster and easier icon finding than the hierarchical menu.
Abstract: In times of major disasters such as hurricane Katrina or the Sichuan earthquake, the need for accurate and timely information is as crucial as is rapid and coherent coordination among the responding humanitarian community. Effective humanitarian information systems that provide timely access to comprehensive, relevant, and reliable information are critical to humanitarian operations. The faster the humanitarian community is able to collect, analyze, disseminate and act on key information, the more effective the response will be, the better needs will be met, and the greater the benefit to the affected populations. This paper presents fundamental principles of humanitarian information management as endorsed by the international humanitarian community, introduces generic systems design premises and presents two recent collaborative efforts in humanitarian information systems development.
Abstract: Multi-criteria decision analysis (MCDA) is a technique for decision support which aims at providing transparent and coherent support for complex decision situations taking into account subjective preferences of the decision makers. However, MCDA does not foresee an analysis of multiple plausible future developments of a given situation. In contrast, scenario-based reasoning (SBR) is frequently used to assess future developments on the longer term. The ability to discuss multiple plausible future developments provides a rationale for strategic plans and actions. Nevertheless, SBR lacks an in-depth performance evaluation of the considered actions.
This paper explores the integration of both techniques that combines their respective strengths as well as their application in environmental crisis management. The proposed methodology is illustrated by an environmental incident example. Future work is to conduct validations on the basis of real-world scenarios by public Dutch and Danish chemical incident crisis management authorities.
Abstract: As language is fundamental to human activities, proficiency in other languages becomes important. Besides for developing abilities for communication, the knowledge is also a tool for a survival. With the introduction of computerized mobile devices, i.e. PDAs, new opportunities for communicating in other language arose. This paper describes a new communication paradigm that is language independent using icon language on a PDA. Users can create iconic messages as realization of their concepts or ideas in mind. The proof of concept tool is able to interpret and convert the messages to (natural language) text and speech in different languages. To provide faster interactions in next icon selection, the tool has icon prediction. Our user test results confirmed that using provided icons our target users could express their concepts and ideas solely using a spatial arrangement of icons.
Abstract: Vigilance concerns the basic human capacity for information processing and is therefore essential to any form of human cognition. Both physical and mental effort are thought to affect vigilance. Mental effort is known for its vigilance declining effects, but the effects of physical effort are less clear. This study investigated whether these two forms of effort affect the EEG (Electro-EncephaloGram; measure of brain activity) and subjective alertness differently. Participants performed a physical task and were subsequently presented with a mental task, or vice versa. Mental effort decreased subjective alertness and increased theta power (i.e. waves with low frequency) in the EEG. Both results suggest a vigilance decline. Physical effort, however, increased subjective alertness and alpha and beta1 power in the EEG. These findings point towards an increase in vigilance. Beta2 power was reduced after physical effort, which may reflect a decrease in active cognitive processing. No transfer effects were found between the effort conditions, suggesting that the effects of mental and physical effort are distinct. It is concluded that mental effort decreases vigilance, whereas physical effort increases vigilance without improving subsequent task performance.
Abstract: Choice of an incorrect representation for the design of automation can dramatically increase
system complexity. Principles from Cognitive Systems Engineering (CSE), which can be used to
identify good representations about the way the ‘world works’, provide a good starting point
for automation design.
This paper discusses, that by choosing the right model for automation design the added
complexity can be limited. But what is the right model for automation? The model of the
environment, or ecology, is preferred above the mental models that human operators have
developed through interacting with the system. The technology has altered the work
environment of the human operator and can have implied to complex or too simplified mental
models. A too complex mental model will bring a too high cognitive load and a simplified
mental model will not be sufficient in all situations. Using the ecology as the basis for the
model of automation, the complexity of the automation is constrained to that of the actual
environment with a minimum share of automation induced complexity.
To illustrate this we considered the design of a conventional autopilot and one based on total
energy control and discuss the mental model pilots have for energy control. Energy control is
the fundamental physics of flight. It is part of the environment thus ecology for pilots and a
proper understanding of energy control helps the pilot to deal with unanticipated event as the
mountain wave condition.
Abstract: Generation algorithms allow for the generation of Virtual Neurons (VNs) from a small set of morphological properties. The set describes the morphological properties of real neurons in terms of statistical descriptors such as the number of branches and segment lengths (among others). The majority of reconstruction algorithms use the observed properties to estimate the parameters of a priori fixed probability distributions in order to construct statistical descriptors that fit well with the observed data. In this article, we present a non-parametric generation algorithm based on kernel density estimators (KDEs). The new algorithm is called KDE-Neuron and has three advantages over parametric reconstruction algorithms: (1) no a priori specifications about the distributions underlying the real data, (2) peculiarities in the biological data will be reflected in the VNs, and (3) ability to reconstruct different cell types. We experimentally generated motor neurons and granule cells, and statistically validated the obtained results. Moreover, we assessed the quality of the prototype data set and observed that our generated neurons are as good as the prototype data in terms of the used statistical descriptors. The opportunities and limitations of data-driven algorithmic reconstruction of neurons are discussed.
Abstract: User demands for usability in mobile context due to the small size of personal data assistants (PDAs) challenge traditional input design. An on-screen keyboard that offers an easier and faster method of entering text with a pen on PDAs, has been developed. We have developed a method for adapting its predictive ability according to user's personal word usage, input context and syntax rules. Frequently used characters are presented to the users in different key sizes and color contrasts according to their relative probabilities to aid visual searching. For this purpose, an experiment has been conducted on which and how to use (user's) data source for faster prediction. In this experiment, we compared four dictionaries recorded from the British National Corpus, personal documents, chat logs and personal e-mails. The experimental results show ways to improve the performance of the word prediction and the language coverage of the word completion.
Abstract: The Dec-POMDP is a model for multi-agent planning under uncertainty that has received increasingly more attention over the recent years. In this work we propose a new heuristic QBG that can be used in various algorithms for Dec-POMDPs and describe differences and similarities with QMDP and QPOMDP. An experimental evaluation shows that, at the price of some computation, QBG gives a consistently tighter upper bound to the maximum value obtainable.
Abstract: Abstract: Reinforcement learning (RL) comprises an array of techniques that learn a control
policy soas to maximize a reward signal. When applied to the control of elevator systems, RL
has the potential of ﬁnding better control policies than classical heuristic, suboptimal policies.
On theother hand, elevator systems oﬀer an interesting benchmark application for the study
of RL. In this paper, RL is applied toa single-elevator system. The mathematical model of
the elevator system is described in detail, making the system easy to re-implement and re-use.
An experimental comparison is made between the performance of the Q-value iteration and
Q-learning RL algorithms, when applied to the elevator system.
Abstract: This research report describes an experiment regarding assessing the effectiveness of actor-agent teams within the SEAT project in the CDM cluster in the ICIS research program. This document focuses on the background of actor-agent teaming, and on a methodology to assess the performance of an actor-agent team in comparison with an actor-only team. The experimental design is described together with the measurements and analysis. The results show that the experimental setup using the REsearch and Simulation toolKit(RESK) provides a repeatable construct. The results of the current performance comparison show no large decrement; but also not a large increment in performance. This is mostly due to the current (low) level ofagent complexity, where improvements are needed in communication capabilities and (more) team-oriented helpful behavior.
Abstract: This white paper contains the report of a first try-out experiment regarding assessing the effectiveness of actor-agent teams within the SEAT project. This document describes the background, experimental design and measurements and analysis of a pilot experiment conducted in July 2008 with an actor-only team.
Abstract: Gas distribution models can provide comprehensive information about a large
number of gas concentration measurements, highlighting, for example, areas of unusual
gas accumulation. They can also help to locate gas sources and to plan where
future measurements should be carried out. Current physical modeling methods,
however, are computationally expensive and not applicable for real world scenarios
with real-time and high resolution demands. This chapter reviews kernel methods
that statistically model gas distribution. Gas measurements are treated as random
variables, and the gas distribution is predicted at unseen locations either using a
kernel density estimation or a kernel regression approach. The resulting statistical apmodels
do not make strong assumptions about the functional form of the gas distribution,
such as the number or locations of gas sources, for example. The major
focus of this chapter is on two-dimensional models that provide estimates for the
means and predictive variances of the distribution. Furthermore, three extensions
to the presented kernel density estimation algorithm are described, which allow to
include wind information, to extend the model to three dimensions, and to reflect
time-dependent changes of the random process that generates the gas distribution
measurements. All methods are discussed based on experimental validation using
real sensor data.
Abstract: Decentralized POMDPs (Dec-POMDPs) are becoming increasingly popular as models for multiagent planning under uncertainty, but solving a Dec-POMDP exactly is known to be an intractable combinatorial optimization problem. In this paper we apply the Cross-Entropy (CE) method, a recently introduced method for combinatorial optimization, to Dec-POMDPs, resulting in a randomized (sampling-based) algorithm for approximately solving Dec-POMDPs. This algorithm operates by sampling
pure policies from an appropriately parametrized stochastic policy, and then evaluates these policies either exactly or approximately in order to define the next stochastic policy to sample from, and so on until
convergence. Experimental results demonstrate that the CE method can search huge spaces efficiently, supporting our claim that combinatorial optimization methods can bring leverage to the approximate solution of Dec-POMDPs.
Abstract: Research on multi-agent systems frequently involves experiments with agents, including situations where humans engage in interactions with agents. Consequently, the ﬁeld of experimental (human) sciences becomes more and more relevant. This paper clariﬁes how things can and often do go wrong in distributed AI experiments. We show the ﬂaws in methodological design in existing literature (both with and without humans) and work out an example involving human test-subjects to introduce the
fundamental issues of experimental design. Furthermore, we provide researchers with an approach to improve their experimental design. We wish to stimulate researchers to conduct better experiments – which will beneﬁt us all.
Abstract: Research suggests opposing alertness effects of mental effort and physical effort: mental effort seems to decrease and physical effort appears to increase subjective alertness. There are indications, however, that physical exercise also leads to feelings of lowered alertness. The well-known multidimensional Thayer alertness scale does not seem to assess physical alertness properly. New items were added to the original scale; these were expected to form a physical factor of subjective alertness. In part 1 of this study, participants filled in the revised Thayer scale before and after a control condition and conditions of physical and mental exercise. Physical exercise only increased feelings of physical fatigue, not of alertness. Mental effort increased feelings of sleepiness. In part 2, a Factor analysis was performed on a larger data set in order to validate the use of a separate physical factor. Indeed, a separate physical factor was found. Besides this physical factor, the analysis revealed the factors “sleepiness”, “calmness”, and “tension”, which have originally been described by Thayer. In conclusion, physical alertness is different from mental alertness. Therefore, an explicit physical factor is required in subjective alertness scales.
Abstract: This paper illustrates objectives of the DIADEM project which focuses on a
novel combination of advanced technologies which facilitate collaborative information
processing in environmental management applications. The emphasis is on the principles
and tools which facilitate creation of a single information space in advanced systems of
systems through a systematic integration of heterogeneous processes. In particular, we
illustrate the main principles of the DIADEM Process Integration framework, which
supports collaborative processing based on a combination of automated reasoning
processes and cognitive capabilities of multiple human experts, each contributing specific
expertise and processing resources.
Abstract: The term vigilance is used frequently in a wide variety of research areas. The British neurologist Sir Henry Head introduced the term to refer to a state of high consciousness. Nowadays, ‘vigilance’ is used in neurophysiological research, but also in the experimental psychological field. Related terms, such as arousal, sustained attention, and tonic alertness are often used jointly with or instead of the term vigilance. It may seem that all these designations can be interchanged freely, but this is not the case. Many investigators differentiate these terms and the distinctions made are not always subtle. The terminological confusion of vigilance involves on the one hand its definition (i.e. clarification of the theoretical construct) with reference to different processes, and on the other hand different measuring procedures. The original definition of vigilance: “… a high state of physiological efficiency” is rather physiological in nature, but the “efficiency” part points to behavior. Head’s clarification that a vigilant state differs from a pure condition of raised excitability appears to be of major importance. The combination of physiological activity and efficient behavior is of great relevance and underlines the difference between vigilance and more basic energetic conditions.
The topic of this paper is the exploration of procedures for measuring vigilance. EEG-measures are very popular and are described first. More specifically, the spectral content of the EEG is investigated. Second, behavioral measures are presented. These concern performance on vigilance tasks. Finally, subjective questionnaires are explored.