Lista de Projetos

Every day a vast amount of misinformation and Fake News are repeated and infinitely shared, reaching millions of people in a short time. The large-scale dissemination of misinformation is one of the major challenges that current societies face, with long-lasting costs to individuals and governments. European Commission’s recent efforts in seeking advice from experts regarding measures to counteract disinformation attest to the urgency of addressing this issue. The fact that people tend to believe in information they repeatedly encounter and to reject claims that contradict what they heard before makes misinformation-correction very difficult. Since most correction strategies entail both a repetition of the false claims and their contradiction, they ironically end up strengthening the validity of the misinformation they attempt to correct. It is thus of the utmost importance to examine the mechanisms that may contribute to the development of effective misinformation-correction actions.

The EU has used the Strategic Energy Technology Plan to transfer power to consumers, by decentralising the energy ecosystem by establishing “100 positive energy districts by 2025 and 80% of electricity consumption to be managed by consumers in 4 out of 5 households”. The SMART-BEEjS recognises that this requires the systemic synergy of the different stakeholders, balancing attention towards technological and policy oriented drivers, citizens and society needs, providers and technology capabilities and value generation system synergies in order to deliver the transition without leaving large parts of the population behind. Smart-BEEjS covers all angles of this eco-system, to train a generation of transformative and influential champions in policy design, techno-economic planning and business model innovation in the energy and efficiency sectors, mindful of the personal and social dimensions, as well as the nexus of interrelation between stakeholders in energy generation, efficiency and management. 

Learning to play music changes brain structure and function, and there is much interest in the idea that these changes might transfer to skills beyond music. Many studies examined if music training improves abilities such as speech and intelligence. However, remarkably little is known about potential transfer effects to social skills, notably the ability to process emotional voices and faces. This effect could be hypothesized from the fundamental link between music and social and emotion processes, and is of central theoretical and applied importance: for understanding brain plasticity, the neurocognitive links between music and socio-emotional abilities, and the potential of music as a therapeutic tool. This project asks if music training improves socio-emotional processing, focusing on three unresolved questions. First, we determine if adult musicians reliably outperform non-musicians at recognizing emotions, and establish the scope of the effect: is it limited to voices, or does it extend to the visual domain (faces)? Is it limited to formally trained individuals,or does it extend to musically sophisticated non-musicians, who developed music skills via informal engagement withmusic? This will clarify previous mixed findings and provide a mechanistic understanding of the effect. A new tool for measuring musical sophistication will be validated and made available to the community. Second, we will combine state-of-the-art magnetic resonance imaging and electrophysiological techniques to delineate the neural mechanisms of the effect. This includes examining how emotions are represented in the trained brain, modulations in the processing time course, and changes in functional connectivity and brain anatomy. This comprehensive approach will add critical new insights into how music drives plasticity. Third, we will conduct a longitudinal study in children to test the effects of a music training program on socio-emotional processing, including pre- and post- training ass...

  It is the objective of the DataSense project to create a computer system that allows acting in the area of the discovery of data considered Sensitive (Sensitive Data Discovery). DataSense has two fundamental objectives that it intends to categorically solve:  • Allow the identification, classification, categorization and relationship of sensitive data present in unstructured information on a large scale in order to allow entities and organizations to obtain an understanding of their sensitive data. • Allow organizations to respond immediately to the content and network (direct and indirect relationships) of the sensitive data they store and process (eg, right to forget)  In order to respond to the aforementioned objectives, DataSense is based on five concepts essential to overcome the state of the art of application and proposes a hybrid architecture that will take the risk of applying the area of Natural Language Processing and Automatic Learning ( Machine Learning) in the critical area of sensitive data protection. The concepts, described in detail in the following chapter are: Sensitive Data (Personal Data), Natural Language Processing, Humanly readable multi-format unstructured information analysis, Intelligence and training supported in human feedback and Interactive Visualization. These basic concepts of the proposed solution are supported by three layers of Artificial Intelligence: Identification of Entities mentioned, Machine Learning models for resolution of Coreferentiation and Feedback and learning of the models