Program

Wednesday, April 14, 2010
09:00 - 10:30

Session 1:
Chair: Malte Helmert

Jan-Georg Smaus
Construction of Buchi Automata for LTL Model Checking Verified in Isabelle/HOL

Abstract: We present the implementation in Isabelle/HOL of a translation of LTL formulae into Buchi automata. In automaton-based model checking, systems are modeled as transition systems, and correctness properties stated as formulae of temporal logic are translated into corresponding automata. An LTL formula is represented by a (generalized) Buchi automaton that accepts precisely those behaviors allowed by the formula. The model checking problem is then reduced to checking language inclusion between the two automata. The automaton construction is thus an essential component of an LTL model checking algorithm. We implemented a standard translation algorithm due to Gerth et al. The correctness and termination of our implementation are proven in Isabelle/HOL, and executable code is generated using the Isabelle/HOL code generator.

Christian Dornhege
Coordinated Exploration with Marsupial Teams of Robots using Temporal Symbolic Planning
Abstract: The problem of autonomously exploring an environment with a team of robots received considerable attention in the past. However, there are relatively few approaches to coordinate teams of robots that are able to deploy and retrieve other robots. Efficiently coordinating the exploration with such marsupial robots requires advanced planning mechanisms that are able to consider symbolic deployment and retrieval actions. In this paper, we propose a novel approach for coordinating the exploration with marsupial robot teams. Our method integrates a temporal symbolic planner that explicitly considers deployment and retrieval actions with a traditional utility-based assignment procedure. Our approach has been implemented and evaluated in several simulated environments and with varying team sizes. The results demonstrate that our proposed method is able to coordinate marsupial teams of robots to efficiently explore unknown environments.

Moritz Gobelbecker
Coming up With Good Excuses: What to do When no Plan Can be Found
Abstract: When using a planner-based agent architecture, many things can go wrong. First and foremost, an agent might fail to execute one of the planned actions for some reasons. Even more annoying, however, is a situation where the agent is incompetent, i.e., unable to come up with a plan. This might be due to the fact that there are principal reasons that prohibit a successful plan or simply because the task's description is incomplete or incorrect. In either case, an explanation for such a failure would be very helpful. We will address this problem and provide a formalization of coming up with excuses for not being able to find a plan. Based on that, we will present an algorithm that is able to find excuses and demonstrate that such excuses can be found in practical settings in reasonable time.

10:30 - 10:45
Coffee Break
10:45 - 12:15

Session 2:
Chair: Daniele Nardi

Luca Iocchi
Multi-Robot Teams for Environmental Monitoring

Abstract: In this talk I will present the achievements of a project aiming at building a mobile reconfigurable multi-robot multi-camera automatic surveillance and monitoring system, that integrates visual and robotic surveillance techniques. I will describe the integrated framework developed to make vision components and robots interact, as well as human behavior analysis based on stereo vision and multi-robot coverage techniques.

Luca Marchetti
Towards a probabilistic approach for Multi-Robot area coverage

Abstract: Video surveillance is an effective application for testing Distributed Data Fusion System. A hybrid architecture, composed by a team of mobile robots and a set of fixed camera, can pursue the surveillance task in a very effective way. In fact, the Multi-Robot System is able to cover a portion of the environment not directly perceived by the cameras, but can also support the belief of fixed sensors. In this talk, we present a probability density function based approach for the problem of Multi-Robot area coverage. The mobile robots are able to autonomously patrol an area, driven by an a-priori knowledge of the environment. The fixed cameras can modify the prior by adding a distributed event handling to the system. The robots can, thus, improve the effectiveness of the surveillance task by supporting the event-driven area coverage. Despite it is still under development, we would like to share some of the preliminary results obtained both in simulated environment and real world applications.

Vittorio Amos Ziparo
Learn to Behave! Rapid Training of Behavior Automata

Abstract: Programming robot or virtual agent behaviors can be a challenging task, and makes attractive the prospect of automatically learning the behaviors from the actions of a human demonstrator. However, learning complex behaviors rapidly from a demonstrator may be difficult if they demand a large number of training samples. We describe an architecture for rapid learning of recurrent behaviors from demonstration. The architecture is based on deterministic hierarchical finite-state automata (HFAs) with classication algorithms taking the place of the state transition function. This architecture allows for task decomposition, statefulness, parameterized features and behaviors, per-behavior feature set customization, and storage of learned behaviors in libraries to be used later on as elements in more complex behaviors. We describe the system, and then illustrate its application in a simple, but nontrivial, foraging task involving multiple behaviors.

12:15 - 13:45
Lunch
13:45 - 15:15

Session 3:
Chair: Jan-Georg Smaus

Malte Helmert
Sound and Complete Landmarks for And/Or Graphs

Abstract: Landmarks for a planning problem are sub goals that are necessarily made true at some point in the execution of any valid plan. Since verifying that a fact is a landmark is PSPACE-complete, earlier approaches have focused on finding landmarks for the delete relaxation $\Pi^+$. Furthermore, some of these approaches have \emph{approximated} this set of landmarks, although it has been shown that the complete set of \emph{causal delete-relaxation landmarks} can be identified in polynomial time by a simple procedure over the relaxed planning graph. Here, we give a declarative characterization of this set of landmarks and show that the procedure computes the landmarks described by our characterization. Building on this, we observe that the procedure can be applied to any delete-relaxation problem and take advantage of a recent compilation of the $m$-relaxation of a problem into a problem with no delete effects to extract landmarks that take into account \emph{delete effects} in the original problem. We demonstrate that this approach finds strictly more causal landmarks than previous approaches and discuss the relationship between increased computational effort and experimental performance, using these landmarks in a recently proposed admissible landmark-counting heuristic.

Robert Mattmuller
Pattern Database Heuristics for Fully Observable Nondeterministic Planning

Abstract: When planning in an uncertain environment, one is often interested in finding a contingent plan that prescribes appropriate actions for all possible states that may be encountered during the execution of the plan. We consider the problem of finding strong and strong cyclic plans for fully observable nondeterministic (FOND) planning problems. The algorithm we choose is LAO*, an informed explicit state search algorithm. We investigate the use of pattern database (PDB) heuristics to guide LAO* towards goal states. To obtain a fully domain-independent planning system, we use an automatic pattern selection procedure that performs local search in the space of pattern collections. The evaluation of our system on the FOND benchmarks of the Uncertainty Part of the International Planning Competition 2008 shows that our approach is competitive with symbolic regression search in terms of problem coverage and speed.

Gabriele Roeger
Stronger Abstractions for the Pancake Problem

Abstract: The pancake problem is a famous search problem where the objective is to sort a sequence of objects (pancakes) through a minimal number of prefix reversals (flips). The best approaches for the problem are based on heuristic search with abstraction (pattern database) heuristics. We present a new class of abstractions for the pancake problem which have three advantages over the approaches considered in previous work. First, our new abstractions are size-independent, ie., do not need to be tailored to a particular instance size of the pancake problem. Second, they are more compact in that they can represent a larger number of pancakes within abstractions of bounded size. Finally, they can exploit symmetries in the problem specification to allow multiple heuristic lookups, significantly improving search performance over a single lookup. Our experimental results show that compared to the pattern databases suggested by Zahavi et al., our new techniques lead to an improvement of one order of magnitude in runtime and up to three orders of magnitude in the number of generated states.

15:15 - 15:30
Coffee Break
15:30 - 17:15

Session 4:
Chair: Reyhan Aydogan

Ehsan Ullah Warriach - Eirini Kaldeli
Composing Services in a Smart Home

Abstract: Domotics is concerned with the automation of the home to increase the safety and comfort of people. We propose a Service-Oriented Computing approach to manage the sensors and actuators of a generic home with the goal of creating an adaptive and interactive middleware. The assumption is that any home device is described as a discoverable service that can be accessed via XML standard descriptions, and a standardized plug-in architecture allows the addition of new devices. The coordination of the home is based on planning techniques that enables the user of the home to command the house issuing high-level instructions, and also allows the home to react to relevant contextual changes. In this talk, we illustrate a platform for smart homes that interacts bidirectional with a visualization/simulation application and the planning techniques that support the dynamic adaptable home coordination.

Elie El-Khoury
Adaptive Middleware for Large-scale Wireless Sensor Networks

Abstract: Wireless Sensor Networks have found more and more applications in a variety of pervasive computing environments. As sensors are becoming more accessible, new challenges arise when deploying in a large-scale environment. The challenges include scalability, network load, communication overhead, security and heterogeneity of the devices. In this presentation, I draw the line over the existing studies on large-scale WSN middleware and then describe the persisting current issues with emphasis on the security challenges. With this in mind, I propose architecture for managing large-scale wireless sensor network, and identify the main research challenges.

Viktoriya Degeler
Searching for strategies to deal with distributed data inconsistencies

Abstract: Pervasive distributed systems must function and make decisions in a constantly changing environment. The information they rely on originates from different distributed sensors and is subject to unbounded delays and faulty and noisy readings. This frequently results in having inconsistent data (outdated, incomplete, or contradictory). The challenge we have is to create an approach for the detection of different kinds of inconsistencies; an approach for reasoning about inconsistencies that we are unable to detect; and a strategy for optimally resolving a task in the presence of contradictory data. In this presentation, we will describe open research questions for dealing with data inconsistencies.

 

 

Thursday, April 15, 2010
09:00 - 10:30

Session 1:
Chair: Eirini Kaldeli

Pavel Bulanov
Requirements and Tools for Variability Management

Abstract: Explicit and software-supported Business Process Management has become the core infrastructure of any medium and large organization that has a need to be efficient and effective. The number of processes of a single organization can be very high, furthermore, they might be very similar, be in need of momentary change, or evolve frequently. If the ad-hoc adaptation and customization of processes is currently the dominant way, clearly it is not the best. In fact, providing tools for supporting the explicit management of variation in processes has a profound impact on the overall life-cycle of processes in organizations. In this presentation, the basic concepts around variability in business process management are defined, and the requirements for having explicit variation handling for (service based) business process systems are considered. Additionally, an evaluation of existing tools for explicit variability management is provided with respect to the requirements identified.

Mahir Can Doganay
Fault tolerance in ubiquitous computing applications.

Abstract: Fault tolerance is of critical importance in pervasive computing scenarios, particularly in domotics applications where system failures can severely disrupt one's daily routine. Ubiquitous computing applications contain numerous small computing and sensing nodes that are crash-prone and can fail due to their battery life. Moreover, these nodes are arranged in an ad-hoc network in which communication between any two nodes is likely to need dynamic routing. Therefore a single node failure could have a bigger effect on the overall system, such as a network partition. In this talk, we present challenges to building failure resilient ubiquitous systems and discuss fault tolerance methods for ubiquitous computing applications.

Sebastian Kupferschmid
Context-Enhanced Directed Model Checking

Abstract: Directed model checking is a well-established technique to efficiently tackle the state explosion problem when the aim is to find error states in concurrent systems. Although directed model checking has proved to be very successful in the past, additional search techniques provide much potential to efficiently handle larger and larger systems. In this work, we propose a novel technique for traversing the state space based on interference contexts. The basic idea is to preferably explore transitions that interfere with previously applied transitions, whereas other transitions are deferred accordingly. Our approach is orthogonal to the model checking process and can be applied to a wide range of search methods. We have implemented our method and empirically evaluated its potential on a range of non-trivial case studies. Compared to standard model checking techniques, we are able to detect subtle bugs with shorter error traces, consuming less memory and time.

10:30 - 10:45
Coffee Break
10:45 - 12:15

Session 2:
Chair: Luca Iocchi

Esra Erdem
Bridging the Gap between High-Level Reasoning and Low-Level Control
We present a logic-based framework to provide robots with high-level reasoning, such as planning and monitoring. This framework uses the action description language C+ to represent actions and changes, and the system CCALC to reason about them. In this approach, the idea is to compute a shortest plan (possibly with concurrent actions) using CCALC, and to monitor its execution in a dynamic environment in such a way to detect and recover from possible failures using both high-level reasoning methods and low-level sensor information. We show the applicability of this framework on two platforms (with LEGO MINDSTORMS NXT robots, and Pantograph robots): we compute a discrete task plan, transform this plan into a continuous trajectory, and monitor its execution to handle collisions.

Matteo Leonetti
Planning and Learning in Partially Observable Domains
Abstract: Automated Planning in highly unpredictable environments is liable to be fragile because of the inevitable inaccuracy of the model and the intrinsic difficulty of dealing with partial observability. We previously developed a method to improve partially specified plans from experience in such challenging environments, not posing any constraint on the source of the plan. In this talk, I will explore the possibility of compiling the information learned during execution back into the theory, so that when the need for replanning arises the new plan can take advantage of previous experience. Such an ability can be deployed together with an execution monitor, being the latter responsible for detecting anomalous situations and the former for trying to avoid such situations altogether in the future. Formalisms in between symbolic and probabilistic reasoning, such as Markov Logic Networks, might provide a sound mechanism to allow for reasoning and learning at the same time.

Feng Xue
Universal Biped Walking Generator With Pattern Feasibility Checking
Abstract: A universal biped walking generator which plans smooth and flexible gaits even on rough terrain is presented in this report. It will not only generate collision-free gaits, but also verify and refine the gaits to make sure they are feasible by the robot when taking different environment and constraints of the mechanism of the robot into account. For the gaits generator, two algorithms are given. The first one is a simplified walking method which only takes a fixed zmp point as the reference in one step. The other one is an advanced ZMP sampling search method which finds a trajectory that will lead to a stable and feasible walking. We then give a sampling based method that verifies and refines the trajectory considering the environment and the constraints of the robot itself.

12:15 - 13:45
Lunch
13:45 - 15:15

Session 3:
Chair: Alexander Lazovik

Jens Witkowski
Truthful Feedback for Sanctioning Reputation Mechanisms
Abstract: For product rating environments, similar to that of Amazon Reviews, it has been shown that the truthful elicitation of feedback is possible through mechanisms which pay buyer reports contingent on the reports of other buyers. We study whether similar mechanisms can be designed for reputation mechanisms at online auction sites where the buyers' experiences are partially determined by a strategic seller. We show that this is impossible for the basic setting. However, introducing a small prior belief that the seller is a cooperative commitment player leads to a payment scheme with a truthful perfect Bayesian equilibrium.

Dapeng Zhang
Constructing Conditional Random Fields Using Switching Attention Learning
Abstract: The feature functions of CRFs are normally specified manually. Learning them from the data is a complex task because the search space grows exponentially with several parameters. We developed a method to solve this problem by using Switching Attention Learning: several learners works one after another to compute a feature function of CRFs; the inputs of a leaner come from the outputs of another, which forms a loop finally; a set of feature functions can thus be obtained by iterating over these learners. A series of experiments were done on several CRFs which are randomly generated. The results show that this method provide a practical way to contruct CRFs from the data.

Dali Sun
Coffee or Tea: Karis at LogiMAT Stuttgart Fair

Abstact: Just sit down, order coffee or tea with a piece of paper. Then, a robot comes and brings the slip. It shan't be long then bring the robot what you want. You don't need to think about the tip because you have no contact with waiter. So is a coffeehouse in the future.
I will give you a talk about our System Karis, which we have demonstrated at Stuttgart LogiMAT Fair. LogiMAT is the International Trade Fair for Distribution, Materials Handling and Information Flow. At this fair, we have successfully demonstrated how the material distribution is working with multiple autonomous robots. We have build a coffeehouse of size 14 * 6 meters and 3 robots was used as waiters to collect the order and bring the goods to the customer. The whole system was running 3 days and 8 hours each day, no Deaklock and no crashing.
Many Techniques have been used on this robotic system, e.g. Mapping, localization, multi-path planning, dynamic obstacle avoiding and autonomous driving. In this Talk the whole system will be at first described. Then our use of multi-path planning and dynamic object avoiding will be described in detail. At the end of the talk is the summary and feature work.

15:15 - 15:30
Coffee Break
15:30 - 17:00

Session 4:
Chair: Tansel Uras

Can Kavaklioglu
Visual Perception Filtering and Adaptation

Abstract: Autonomous robots can initiate their mission plans only after gathering information about the environment. Therefore reliable perception information plays a major role in the overall success of task accomplishment. Essentially most of the errors in the lower level perception methods can be traced to sensor inaccuracies. One way of making perception information more reliable is using filters to compensate for errors in low level perception methods. Another approach is trying to achieve better adaptation to environment changes.

Baris Gokce  
Parameter Optimization for a Signal-Based Omni-Directional Biped Locomotion by Using Evolutionary Strategies

Abstract: The ultimate goal of RoboCup depends heavily on the advances in the development of humanoid robots. Flexible walking is a crucial part of playing soccer and bipedal walking has been a very active research topic in RoboCup. In this paper a signal-based omni-directional walking algorithm for Aldebaran Nao humanoid robot is presented. Inspired from the existing methods in the literature, the proposed method models the omni-directional motion as the combination of a set of periodic signals. The parameters controlling the characteristics of the signals are encoded into genes and Evolutionary Strategies is used to learn an optimal set of parameters.

Reyhan Aydogan
Learning Preferences in Automated Negotiation

Abstract: In e-commerce, consumers and producers interact with each other to accomplish their goals. During this interaction, they need to have a consensus on the service since their preferences may conflict. In this setting, representation of participants' preferences and reasoning on these preferences play a key role. To generate well-targeted offers, an agent should not only consider its own preferences but also other participant's preferences. However, in general it may not be possible to know other participants' preferences. Thus, learning others' preferences from the bids exchanged during the negotiation becomes an important task. To achieve this, we have purposed an inductive learning algorithm, namely Revisable Candidate Elimination Algorithm (RCEA) that can be used by a producer during negotiation to learn consumer's preferences in the form of conjunctive and disjunctive constraints.




Friday, April 16, 2010
09:00 - 10:30

Session 1:
Chair: Marco Aiello

Michael Brenner
Creating Dynamic Story Plots with Continual Multiagent Planning

Abstract: An AI system that is to create a story (autonomously or in interaction with human users) requires capabilities from many subfields of AI in order to create characters that themselves appear to act intelligently and believably in a coherent story world. Specifically, the system must be able to reason about the physical actions and verbal interactions of the characters as well as their perceptions of the world. Furthermore it must make the characters act believably--i.e. in a goal-directed yet emotionally plausible fashion.  Finally, it must cope with (and embrace!) the dynamics of a multiagent environment where beliefs, sentiments, and goals may change during the course of a story and where plans are thwarted, adapted and dropped all the time.  In this talk, we describe a representational and algorithmic framework for modelling such dynamic story worlds, Continual Multiagent Planning. It combines continual planning (i.e. an integrated approach to planning and execution) with a rich description language for modelling epistemic and affective states, desires and intentions, sensing and communication. Analysing story examples generated by our implemented system we show the benefits of such an integrated approach for dynamic plot generation.

Daniele Calisi
Towards a unifying framework for wheeled robot motion systems
Abstract: The proliferation of different approaches for robot motion has many advantages: as a complete and satisfying solution does (still) not exist, many research lines can be explored in order to achieve a better understanding of the problem and its possible solutions, depending on the application at hand. Nonetheless, many methods have been used only to solve speci?c problems and their strengths and generality sometimes is far from being completely understood. As a matter of fact, a general framework for robot motion systems is still missing, where different methods and their variations can be arranged, composed and compared systematically. In this talk, we give a characterization of the local motion methods and techniques that allows for the seamless integration of pure-reactive behaviors and local planners, taking advantage of both techniques. While the former are more suitable for dynamic environments and high speed motions, the latter are usually more accurate and can cope with a larger set of situations and tasks: we think that future motion systems should autonomously trade speed for safety or accuracy, depending upon the context and the specific task to accomplish, adapting or choosing among different methodologies.

Gabriele Randelli
Novel human-robot interaction paradigms through tangible user interfaces

Abstract: Tangible user interfaces (TUIs) promise to ease human-robot interaction in manifold activities. In order to identify their potential benefits, it is important to assess how they influence human spatial cognitive processes with respect to more traditional interfaces. In this talk, we present an interaction framework that correlates human cognitive processes to novel human-robot interaction paradigms. We also show some relevant robotic applications in which we adopted TUIs, implementing such paradigms. Finally, we introduce the experiment context we are going to attend, in order to evaluate some interesting considerations about TUIs in robotics.

10:30 - 10:45
Coffee Break
10:45 - 12:15

Session 2:
Chair: Gabriele Roeger

Tansel Uras
Genome Rearrangement and Planning: Revisited
Evolutionary trees of species can be reconstructed by pairwise comparison of their entire genomes. Such a comparison can be quantified by determining the number of events that change the order of genes in a genome. Earlier Erdem and Tillier formulated the pairwise comparison of entire genomes as the problem of planning rearrangement events that transform one genome to the other. We reformulate this problem as a planning problem to extend its applicability to genomes with multiple copies of genes and with unequal gene content, and illustrate its applicability and effectiveness on three real datasets: mitochondrial genomes of Metazoa, chloroplast genomes of Campanulaceae, chloroplast genomes of various land plants and green algae.

Duygu Cakmak
Reconstructing phylogenies from large data sets using ASP

We introduce a novel method for reconstructing large phylogenies for a set of species, using Answer Set Programming(ASP). Our method takes into account the shared traits of the species specified by the experts, the domain-specific information about the well-known subgroups of the species, and some weight measures to quantify the plausibility of phylogenies. Then it reconstructs a phylogeny for the given species in two steps: first it reconstructs a phylogeny for each specified subgroup of species, and then it reconstructs a main phylogeny that combines all the subgroups. At each step, our method tries to find a max-weighted phylogeny with a large number of traits that are compatible with the phylogeny (i.e., that can explain the evolution of the species from their common ancestor). Based on this ASP-based method, we have implemented aphylogenetics system, called PHYLORECONSTRUCT-ASP, and used it to reconstruct the evolutionary history of the genus Quercus (oak trees) based on the morphology of acorns. We have observed that all the phylogenies computed by PHYLORECONSTRUCT-ASP are plausible from the biology viewpoint, suggesting the appropriateness of the compatibility criterion and the weight measures; furthermore, most of the computed phylogenies are different from the existing ones, suggesting a new perspective on the evolution of oak trees.

Halit Erdogan
Finding Similar or Diverse Solutions in Answer Set Programming
Abstract: Although, for many computational problems, the main concern is to find a best solution (e.g., a most preferred product configuration, a shortest plan, a most parsimonious phylogeny), for some problems, computing a subset of good solutions that are diverse or similar may be desirable. For instance, in product configuration, one could be interested in obtaining several diverse configurations of a product instead of checking all possible configurations, to pick one. In planning, it may be desirable to compute a set of plans that are similar to each other, so that, when the plan that is being executed fails, one can switch to a most similar one. Motivated by such applications, we study the problem of computing similar or diverse solutions in answer set programming (ASP), and then show the applicability of our approach to another interesting problem: phylogeny reconstruction (i.e., computing leaf-labeled trees, called phylogenies, to model the evolutionary history of a set of species).

12:15 - 13:45
Lunch
13:45 - 15:15

Session 3:
Chair: Halit Erdogan

Matthias Westphal and Stefan Wolfl
Guiding Manipulation Plan Generation by Qualitative Spatial Reasoning (joint work with Christian Dornhege)

Abstract: Motion planning is a complex task for a robot that needs to grasp objects in its environment, specifically if narrow spatial conditions restrict the action space of the robot arm, which is equipped with a gripper. Usually probabilistic roadmap planners are used to generate such plans, but plans from such planners often lead to arm movements that are sub-optimal and also far from how a human might perform a comparable grasping task. In our talks we present a hybrid framework that improves the quality of generated plans in many spatial situations. The idea is to guide the probabilistic roadmap planner by a qualitative spatial plan that provides an approximation of a geometric solution on a qualitative level of description.
In the first talk we state the problem on a formal level and discuss how the problem can be cast as a benchmark problem for research in the qualitative spatial reasoning domain. Moreover, we discuss two simple qualitative formalisms in which a baseline solution to the problem can be represented, and demonstrate how qualitative plans can be employed to guide the geometric planning process.
In the second talk we describe the overall architecture of our hybrid planning approach. We present the results of an evaluation of the architecture in different spatial settings and discuss how the quality of generated plans can be improved by qualitative spatial reasoning techniques. An outlook on further possible improvements completes our talks.

Patrick Eyerich and Thomas Keller
High-Quality Policies for the Canadian Traveler's Problem

Abstract: We consider the stochastic variant of the Canadian Traveler's Problem, a path planning problem where adverse weather can cause some roads to be untraversable. The agent does not initially know which roads can be used. However, it knows a probability distribution for the weather, and it can observe the status of roads incident to its location. The objective is to find a policy with low expected travel cost.
We introduce and compare several algorithms for the stochastic CTP. Unlike the optimistic approach most commonly considered in the literature, the new approaches we propose take uncertainty into account explicitly. We show that this property enables them to generate policies of much higher quality than the optimistic one, both theoretically and experimentally.

15:15 - 15:30
Coffee Break
15:30 - 17:00

Session 4:
Chair: Stefan Woelfl

Akin Gunay
Service Matchmaking Revisited: An Approach based on Model Checking

Abstract: The aim of service discovery is to find services that satisfy user requests in a precise and efficient manner. An important aspect of service discovery is service matchmaking, which constitutes the mechanism to map appropriate services to requests. Current service matchmaking approaches mainly use the knowledge about the interface descriptions of services. However, these approaches suffer from lack of precision since they do not consider the internal processes of services.
This paper proposes a novel service matchmaking approach that uses the internal process models of services as primary source of knowledge. To reason about the internal process models and to identify matching services to requests, we use model checking as a reasoning mechanism. In order to facilitate partial matches, we use ontologies and relaxation techniques to generate alternative requests. Hence, even when a request cannot be satisfied by a service, our approach can identify which similar requests are satisfied by the service. This important information can enable better service selection for the service consumers. We also provide a guideline to illustrate how our proposed matchmaking approach can be realized using recent technologies from Web services and formal verification domains in a real world setting.

Andrea Pagani
A peer-to-peer approach to energy generation and distribution: a preliminary study

Abstract: The energy sector is undergoing important changes that will impact the way users access and consume energy. With an increasing unbundling process affecting the energy market, anybody can at the same time be energy producer and consumer, thus giving the possibility to sell and buy energy on a free market. The shift towards distributed generation has an impact on the power grid, especially at the medium and low voltage segments, both from the physical point of view and from the information management perspective. The talk outlines the current situation of the power grid and illustrates a scenario where micro-generation and Smart meters have a paramount role. Furthermore, we look at the layout of the power grid by performing a complex network analysis of the North Netherlands network for which we present initial findings.

Ando Emerencia
Intelligent Web Applications To Support People With Schizophrenia

Abstract: There is a paradigm shift in modern day health care towards a patient centered approach, where patients are in control of managing their disease, often through the use of Web applications. For people suffering from schizophrenia however, little has been done so far and this is often attributed to the fact that these people have different needs with respect to the structure, content and user interface of a Web site. This is why we work in cooperation with researchers from the University Medical Center Groningen (UMCG) to design an intelligent web application specifically for this group of patients. By data mining patients' test results, questionnaire answers and demographic information from electronic medical records, we aim to construct user profiles that model a patient's state (e.g. problems, satisfaction, needs, preferences) using a hierarchical ontology that relates the concepts using domain knowledge. Using these profiles, relevant information and intelligent suggestions can be offered, personalized and localized for each patient (e.g. recommend a local sports facility to a patient that wants to lose weight). Through unobtrusive monitoring (e.g. navigational paths, links (not) clicked, time spent reading, text scrolled over) and collaborative filtering, we can update these profiles or estimate missing values. For future interoperability with personal health record services (e.g. import/export data to Google Health, Microsoft HealthVault) it is important to adhere to international standards for terminology (SNOMED-CT, LOINC), communication (HL7, ISO13606) and patient summaries (CDA, CCR). In this talk I will discuss schizophrenia, give a brief overview of artificial intelligence approaches in medicine and present an initial design for the Web application.