Abstract
Deep Transfer Learning (DTL) emerged as a new paradigm in machine learning in which a deep model is trained on a source task and the knowledge acquired is then totally or partially transferred to help in solving a target task. Even though DTL offers a greater flexibility in extracting high-level features and enabling feature transference from a source to a target task, the DTL solution might get stuck at local minima leading to performance degradation-negative transference-, similar to what happens in the classical machine learning approach. In this paper, we propose the Source-Target-Source (STS) methodology to reduce the impact of negative transference, by iteratively switching between source and target tasks in the training process. The results show the effectiveness of such approach.

Abstract
Robots created for education have different purposes, from learning algorithms to learning robotics. It has been shown that robots can increase the student's interest. Our proposed robot, REDi, uses wire programming to introduce students to the basics of robotics and algorithms. With this robot, attractive, interactive demonstrations can be achieved even with students that have no background in the area.

Abstract
As the amount of content and the number of users in social relationships is continually growing in the Internet, resource sharing and access policy management is difficult, time-consuming and error-prone. In order to aid users in the resource-sharing process, the adoption of an entity that recommends users with access policies for their resources is proposed, by the analysis of (i) resource content, (ii) user preferences, (iii) users' social networks, (iv) semantic information, (v) user feedback about recommendation actions and (vi) provenance/ traceability information gathered from action sensors. A hybrid recommendation engine capable of performing collaborative-filtering was adopted and enhanced to use semantic information. Such recommendation engine translates user and resources' semantic information and aggregates those with other content, using a collaborative filtering technique. Recommendation of access policies over resources promotes the discovery of known-unknown and unknown-unknown resources to other users that could not even know about the existence of such resources. Evaluation to such recommender system is performed.

Abstract
The integration of ICT technologies will contribute to meet the objectives of smart grids by applying intelligent strategies to implement their automation functions. The use of the real electrical power energy infrastructure is hardly to be managed. The use of simulation constitutes an alternative to model and test in a simple way these complex power electrical grids, improving the system's reliability, resilience and stability. However, power simulators don't reflect the unpredictable behaviours caused by the collaborative interference of intelligent and cognitive systems, being required an easy and transparent platform to interconnect the intelligent control with the power emulator platforms. This paper introduces a two-layer framework integrating multi-agent-based control systems with power simulator tools, allowing to create a virtual environment to test and simulate the different control strategies and helping the deployment of the future smart grid concept. An instantiation of this framework is deeply analysed for the use of GridLab-D as power simulation tool.

Abstract
This paper presents the study of modeling a quality supplier management system, based on a particular case of a Portuguese company in the auto industry that supplies parts to the major automobile manufacturers. For this, a study of a practical case of Supplier Quality Management (SQM) that follow the ISTO/TS 16949 standard was done, to analyze the functioning of their processes. Finally, a model is proposed streamline and automate supplier performance management processes, with the aim of improving the efficiency throughout the supplier chain.

Abstract
Wireless Sensor Networks, in particular Wireless Body Area Networks, is a technology suggested by the research community as allowing elderly people, or people with some kind of disability, to live in a safer, responsive and comfortable environment while at their homes. One of the most active threats to the autonomous life of blind people is the quantity and variety of obstacles they face while moving, whether they are obstacles in the footpath or obstacles coming out from the walls of buildings. Hence, it is necessary to develop a solution that helps or assists blind people while moving either in indoor or outdoor scenarios, simultaneously allowing the use of the use of white cane or the Seeing Eye dog. In this article, the authors propose the use of an ultra-sound based body area network for obstacle detection and warning as a complementary and effective solution for aiding blind people when moving from place to place. According to the cost estimates of the solution and to the negligible setup time, this could be a real effective complementary solution for blind people. (c) 2015 The Authors. Published by Elsevier B.V.