Abstract
This paper presents two contributions: (i) a system architecture capable of staging platform-independent choreographies within different virtual worlds, and (ii) an ontology-based solution for capturing and representing multi-user choreographies with reduced time/effort. We argue that choreographies for virtual worlds should be clearly separated from the technical characteristics of their execution in virtual world technological platforms. Due to the heterogeneity of the various virtual worlds and their domain requirements, we propose exploiting the modularity, generality, and granularity dimensions of ontologies to simplify and empower the choreography modeling capabilities. Instead of a unique ontology, several ontologies with different levels of generality and granularity can be progressively combined to support the modeling requirements of a given choreography. Because these ontologies are aligned with the ontology of each specific virtual world platform, the mapping and transformation between the core ontology is simplified and automated, thus reducing the development and time-to-market.

Abstract
The introduction of Adaptive Cruise Control (ACC) could be very helpful for making the longitudinal driving task more comfortable for the drivers and, as a consequence, it could have a global beneficial effect on road safety. However, before or during the usage of the device, due to several reasons, drivers might generate in their mind incomplete or flawed mental representations about the fundamental operation principles of ACC; hence, the resulting usage of the device might be improper, negatively affecting the human-machine interaction and cooperation and, in some cases, leading to negative behavioural adaptations to the system that might neutralise the desirable positive effects on road safety. Within this context, this paper will introduce the methodology which has been developed in order to analyse in detail the topic and foresee, in the future, adequate actions for the recovery of inaccurate mental representations of the system.

Abstract

Abstract
Schematic Maps are mainly used for depicting transportation networks. They are generated through a schematization process where irrelevant details are eliminated and important details are emphasized. This process, being manually performed by teams of expert designers, is expensive and time consuming. Such manual execution is unsuitable for the production of schematic maps for location-based services or on-demand schematic maps, as near real-time and user-centered properties are needed. This work proposes GeneX, a framework that can support the automated generation of schematic maps. The framework and a new algorithms developed were able to completely eliminate erroneous map point placement, and to decrease by 33% the contention for map point placement, producing schematic maps without human intervention in soft real time.

Abstract

Abstract
Several predictive systems are nowadays vital for operations and decision support. The quality of these systems is most of the time defined by their average accuracy which has low or no information at all about the estimated error of each individual prediction. In these cases, users should be allowed to associate a measure of reliability to each prediction. However, with the advent of data streams, batch state-of-the-art reliability estimates need to be redefined. In this chapter we adapt and evaluate five empirical measures for online reliability estimation of individual predictions: similarity-based (k-NN) error, local sensitivity (bias and variance) and online bagging predictions (bias and variance). Evaluation is performed with a neural network base model on two different problems, with results showing that online bagging and k-NN estimates are consistently correlated with the error of the base model. Furthermore, we propose an approach for correcting individual predictions based on the CNK reliability estimate. Evaluation is done on a real-world problem (prediction of the electricity load for a selected European geographical region), using two different regression models: neural network and the k nearest neighbors algorithm. Comparison is performed with corrections based on the Kalman filter. The results show that our method performs better than the Kalman filter, significantly improving the original predictions to more accurate values.