Search

The past few decades have seen important efforts to introduce more ethics into engineering education programs around the world, while adapting to the evolution of ethical concerns, notably in relation to digital technology and artificial intelligence. Even if pedagogical approaches based on the use of case studies or code of ethics remain among the most popular, other less well known techniques such as game-based approaches have also been identified as effective. The goal of this workshop is to offer participants an opportunity to explore how games can be used as learning experiences to develop students’ ethical knowledge and skills. Participants will first get to play an online game which focuses on ethical issues in the domain of artificial intelligence, before reflecting on their experience and discussing the potential of game-based approaches for engineering ethics education.

International audience; This paper introduces an approach for fault detection in ambient-intelligent environments. It proposes to compute predictions for sensor values, to be compared with actual values. As ambient environments are highly dynamic, one cannot pre-determine a prediction method. Therefore, our approach relies on (a) the modeling of sensors, actuators and physical effects that link them, and (b) the automatic construction at run-time of a heterogeneous prediction model. The prediction model can then be executed on a heterogeneous modeling platform such as ModHel'X, which yields predicted sensor values.

In the context of Model Driven Engineering, the use of multiple modeling paradigms for developing complex systems is both unavoidable and essential. It results in the heterogeneity of the models representing the considered system and makes global reasoning about the system difficult. The objective of multi-paradigm modeling is to ease the joint use of heterogeneous models during the development cycle. In the work presented in this dissertation, we focus on the study of the heterogeneity of models and propose an approach to multi-paradigm modeling. We first qualify the causes of the heterogeneity of models with respect to the development cycle and we identify several types of heterogeneity. On this basis, we propose a framework for the study of the domain of multi-paradigm modeling with several research axis. The multidisciplinarity of multi-paradigm modeling allows the use of techniques from various fields. We propose a survey and a classification of the techniques which are relevant with respect to heterogeneity. The range of the techniques that we present includes model transformation, meta-model composition, model composition, component adaptation, co-simulation and megamodels. Then we present ModHel'X, the approach to the composition of models for multi-paradigm modeling that we developed. It relies on the concept of model of computation and allows: 1. the specification of the semantics of a modeling language in an executable way by specializing an abstract semantics for models of computation that we developed; 2. the explicit specification of the composition mechanism between heterogeneous models through a special modeling structure called interface block; 3. the simulation of the global behavior of heterogeneous models thanks to a generic execution algorithm that we defined. ModHel'X has been implemented in a framework based on EMF (Eclipse Modeling Framework)., Dans le contexte de l'Ingénierie Dirigée par les Modèles, l'utilisation de multiples paradigmes de modélisation pour développer un système complexe est à la fois inévitable et essentielle. Les modèles qui représentent un tel système sont donc hétérogènes, ce qui rend tout raisonnement global sur le système difficile. L'objectif de la modélisation multi-paradigme est de faciliter l'utilisation conjointe de modèles hétérogènes pendant le cycle de développement. Les travaux exposés dans cette thèse concernent l'étude de l'hétérogénéité des modèles et la conception d'une approche pour la modélisation multi-paradigme des systèmes. Nous caractérisons les causes de l'hétérogénéité des modèles par rapport au cycle de développement puis identifions différents types d'hétérogénéité. En nous basant sur ces causes d'hétérogénéité, nous proposons un cadre d'étude pour le domaine de la modélisation multi-paradigme avec différents axes de recherche. La multidisciplinarité de la modélisation multi-paradigme rend applicables des techniques issues de différents domaines. Nous proposons un état de l'art et une classification des techniques dont nous avons étudié la pertinence par rapport à l'hétérogénéité. La gamme des techniques présentées inclut les transformations de modèles, la composition de méta-modèles, la composition de modèles, l'adaptation de composants, la co-simulation ou encore les méga-modèles. Nous présentons ensuite ModHel'X, l'approche de composition de modèles pour la modélisation multi-paradigme que nous avons développée. Elle s'appuie sur le concept de modèle de calcul et permet : 1. de spécifier la sémantique d'un langage de modélisation de manière exécutable à travers la spécialisation opérationnelle d'une sémantique abstraite pour les modèles de calcul ; 2. de spécifier explicitement les mécanismes de composition à utiliser entre des modèles hétérogènes via une structure de modélisation appelée bloc d'interface ; 3. de simuler le comportement global de modèles hétérogènes par un algorithme générique d'exécution que nous avons défini. Une implémentation de ModHel'X a été réalisée sous la forme d'un framework s'appuyant sur EMF (Eclipse Modeling Framework).

The use of digital tools has drastically increased in engineering education, accelerated by the COVID-19 pandemic. These tools generate important ethical issues, in particular in terms of privacy and fairness. However, very few teacher training programmes address those topics, which means that teachers are often left to figure out by themselves how to address these issues when they want (or have) to use digital tools in their teaching. In this workshop, participants will be introduced to a pragmatic approach to the ethical design of learning experiences that involve digital tools using a visual thinking guide called a ‘canvas’. Applied and hands-on, this workshop will help participants to develop a practical understanding of the specific ethical issues related to the use of digital tools in teaching and to integrate ethical reflection into design processes when digital technology is involved.

This paper presents the results of a project aimed to measure changes in attitudes and perception of science, engineering and architecture first year students about working in a professionally diverse environment. At the beginning of the Global Issues course of the autumn semester in 2014 all 1, 800 students were asked to complete two psychometric tests: The Readiness for Inter-professional Learning Scale (RIPLS) aims at measuring attitudes towards inter-professional learning (McFayden, AK, et al, 2005); and the Engineering and Science Issues Test (ESIT) aiming to measure moral reasoning in engineering by responding to professional ethical dilemmas. Students were asked to repeat the task 13 weeks later after having followed lectures and completed the requested interdisciplinary group work (Holzer, A., Vonèche-Cardia, I., et al, 2016). A superficial analysis of the RIPLS scores shows an overall decreased positive attitude towards interdisciplinary work. A more detailed analysis suggest a paradox: While students agree that team work skills will make them better professionals they disagree on the benefit of practicing such skills before qualification. The paper will focus in discussing the reasons for such paradox looking into the timing of the testing (too early), students' disciplinary identity, and other potential reasons such as conflicts in schedules and difficulties in communication. Our results are taken as a clear sign of the need to keep in mind the ‘hidden curriculum’ in scientific and engineering education when tailoring the learning activities supporting interdisciplinary courses (Tormey, R, Le Duc, I, et al, 2015).