Lista de Projetos

The AgroPro project aspires to provide the agriculture professionals acquaintance with the available airborne technology & software as well as with UAV international legislation & regulations. The acquaintance with mission planning & the available open source and commercial tools.

The aim of the SmartVitiNet project is to (a) scale-up, pilot and bring to the market an innovative holistic phytosanitary and plant protection system based on the use of unmanned aerial vehicles, new observational platforms and new ready to use sensors, and (b) establish a Competence Center for Precision Viticulture. The proposed research will utilize complementary knowledge, experiences and infrastructure of all partners to achieve the proposed innovative results. The sustainability of the undertaking will be ensured thanks to the establishment of the Competence Centre for Precision Viticulture which aims to upskill sector professionals, create expert networks, facilitate permanent flows of knowledge transfer between academia, innovative SMEs, viticulture professionals and regional authorities to increase sector competitiveness, while enacting EU environmental policies, reducing sector health impact and risks of food pollution.

EMPOWER will focus on education for children with neurodevelopmental disorders (NDDs). Children with NDDs can experience difficulties with language and speech, motor skills, behaviour, memory, learning, or other neurological functions. Technological solutions that can respond to such individual needs have the potential to both improve the quality and inclusiveness of the education of these children and support teachers in carrying out their educational vocation. From a technological perspective, the challenge is not only to deliver the resulting educational program but also to do so accurately and to the benefit of the child. From an ethical perspective, several challenges come together in the trade-off between the potential educational benefits and the necessity to process relevant information regarding the children via measurements and algorithms that shape the educational program. In the proposed AI regulations of the EU (Artificial Intelligence Act, EC/2021), this is a high- risk endeavour. Together, this application domain is therefore a challenging one in that it unites sensitive cases of the obstacles one is likely to encounter in digitizing education. Addressing these challenges is therefore also an opportunity to shed more light on the future of technology and AI in education as the ability to address these challenges in their extreme form will lead to insights that are relevant more generally.

A modelação Bayesiana de equações estruturais (BSEM) tem vindo a receber crescente interesse principalmente devido à sua capacidade de resolver alguns dos problemas encontrados na abordagem frequencista convencional (por exemplo, não convergência, casos de Heywood, tamanho da amostra, soluções inadmissíveis) e porque permite ajustar modelos complexos que os métodos clássicos de máxima verossimilhança podem ter dificuldades para estimar (Merkle & Rosseel, 2018). No entanto, a BSEM pode ser computacionalmente intensiva. De facto, o custo computacional da análise Bayesiana prejudicou o uso mais frequente da estatística Bayesiana. O uso de métodos Bayesianos tem melhorado, principalmente devido aos avanços computacionais, fornecendo aos investigadores ferramentas mais flexíveis e poderosas. Atualmente, a análise Bayesiana é um ramo estabelecido da metodologia para estimação de modelos (van de Schoot et al., 2021). Em parte, isso deve-se a dois aspetos: o aumento da popularidade da metodologia Bayesiana e o advento dos métodos Monte Carlo via cadeias de Markov (MCMC; Depaoli, 2021). A estatística Bayesiana beneficia dos métodos MCMC, pois a análise Bayesiana depende muito da integração multidimensional. O MCMC compreende um conjunto de algoritmos computacionais que podem ajudar a resolver situações de modelagem complexas e de elevada dimensionalidade (South et al., 2022). O MCMC pode ajudar a estatística Bayesianas — por exemplo — reconstruindo a distribuição posterior (Depaoli, 2021). O MCMC pode beneficiar muito de um ambiente de computação paralelo, que permita realizar cálculos extensos simultaneamente. Os avanços no hardware do computador de consumo tornam a computação paralela amplamente disponível para a maioria dos utilizadores. Muitas placas gráficas de vídeo de computador suportam computação paralela. O uso das unidades de processamento gráfico (GPU) geralmente proporcionam ganhos significativos em termos de desempenho (Češnovar et al., 2019). Não há dúvida...

Despite the Union’s effort to fight against online hate speech (OHS), several reports showed an increase in OHS during 2020-21. The current pandemic provided a context for increased scapegoating and stigmatization, and minority groups are disproportionally targets of hatred discourse. OHS is a persistent threat to the Union’s values and there is a need for more knowledge on its content, detection and countering, as highlighted in the current Call. Portugal, as other member states, has seen an escalation of hate speech against immigrants, racial/ethnic groups, and LGBTIQ communities. However, there is no systematized knowledge nor tools designed to detect, monitor and prevent OHS against these communities. Our project aims at addressing this need, offering a comprehensive, participatory and culturally sensitive approach to analyse, detect, and counter, direct and indirect OHS in Portuguese language.