Research

Food and nutrition security and sustainable agriculture (FNSSA) are among the critical development concerns in Africa. FNSSA is among the policy priorities in many African countries and at the center of the AU/EU international development agenda. FNSSA was the priority R&I area in the AU-EU High Level Policy Dialogue on Science, Technology, and Innovation. This was further substantiated by the establishment of the Pan-African Network for Economic Analysis of Policies (PANAP). StEPPFoS thus, aims contribute to theFNSSA 10-year road map and the global transition towards sustainable food systems through the implementation of activities that link PANAP to the FNSSA partnership. Specific objectives are to improved capacities of stakeholders, to enhanced science-policy interface, to improve strategies that promote scientific support within policy development, to expand and strengthen the PANAP Network. StEPPFoS will be implemented through its 8 work-packages over a period of 48 months. Capacity building,stakeholder engagements, participatory monitoring, evaluation, and learning are the main methods to be used to deliver StEPPFoS objectives whiles adhering to open science principles and effective data management practices. The consortium is well positioned to deliver these objectives as it draws on the expertise and experiences of its partners drawn from both Europe and Africa credible academic, research, and policy institutions.

LEAP aims to equip learners with in-demand ICT skills, bridging the gap between industry and academia by offering a flexible and inclusive learning approach. The project aims to develop and offer learning content that is efficient, inclusive and pertinent to the current state of the ICT labour market. The curriculum focuses on complex systems, statistics, data analytics and machine learning, with an experiential pedagogical approach that fosters transferable and forward-looking skills.

LEAP aims at bridging the gap between industry and academia by offering flexible and inclusive outcomes-oriented and student-oriented learning, integrating Complex Systems, Data Analytics, Statistics, and Machine Learning. The planned activities follow the ADDIE model, including needs analysis, pedagogy research, course design, and evaluation. Dissemination activities enhance project impact.

The Narratives Observatory to combat Disinformation in Europe Systemically (NODES) has been designed to deploy the power of narratives in the fight against disinformation within the European public sphere. The objective of NODES is to develop and critically assess a novel large-scale mechanism to analyse and monitor the origin, development and spread of narratives classified as “disinformation”, to effectively provide quantitative and qualitative data on the narrative approach to disinformation, including the implementation of the NODES Narrative Methodology, recommendations for future policy action and novel practices for practitioners in the field.

The complexity of the contemporary, multi-issue, multi-lingual, multi-platform and multi-cultural European public sphere will be reflected through the methodology being applied to three different case studies (COVID19, Climate Change and Migration) and four different languages (English, French, Polish and Spanish) in order to be able to analyse the topology of network structures within the public sphere from a narrative perspective.

Bringing together the experience and expertise within the consortium that represent leading actors across the interdisciplinary and intersectoral spectrum of stakeholders involved in the issue, NODES has been designed to be able to be operative from M1 by building on the tools and methodologies already developed by and available to the partners. Thus benefiting from the experience, expertise and know-how developed over decades.

Impact is a key feature of the success of the project as it aims to provide a concrete and systematic contribution to combatting disinformation and strengthening our democracies. Building on the diverse outreach and expertise of the consortium, NODES will draw conclusions, propose measures and develop recommendations for the different audiences addressed by this concern including: Policy-makers, Practitioners, Academia, Journalists/Media, General audience.

Project website: https://nodes.eu/

ELISE aims to make Europe competitive in AI through a network of excellence. The best European researchers in machine learning and AI have worked together to attract talent, to foster research through collaboration, and to inspire and be inspired by industry and society. While ELISE starts from machine learning as the current most prominent method of AI, the network invites in all ways of reasoning, considering all types of data, applicable for almost all sectors of science and industry. While being aware of data safety and security, and while striving to explainable and trustworthy outcomes we aim to create a force to Europe.

ELISE will run a PhD student and a postdoc programme to attract and to educate world-class talents to Europe. It will operate a Fellows programme for groundbreaking research and high-profile workshops to develop AI methods applications further. Industry involvement is guaranteed by the many connections members of ELISE have with industry, on average one for every member and one start-up for every second member of ELISE.ELISE will demonstrate a fraction of their research in use cases to be implemented in AI4EU and the SME’s of Europe. Additional impact will be created to SME’s through open calls. The current practice of ELISE members of spin-off research in SME’s once a break-through is achieved will be stimulated through incubators.The current practice of participating in dissemination and debate that many members of ELISE are used to will be continued to develop a mature acceptance of AI throughout Europe for the benefit of all and in cooperation with all.

ELISE is built on 105 organisations in total, in which the 202 core contributors have actively indicated they will help build and profit from the networks of PhD-students and scholars.
ELISE includes 60 ERC grants of their active supporters. By their citation and other accepted scores of scientific quality, ELISE is the network that combines in Europe excellence in AI.

UNIVE-ECLT, as Linked-Third Party of the Consorzio Interuniversitario Nazionale per l’Informatica (CINI), will contribute to the realization of the following deliverables:

• D6.2: ELISE challenges accessible and showcased on the AI4EU platform (M36)
• D6.4: Booklet and media content on ELISE dissemination activities to the large public and industries (M24)
• D6.5: Booklet and media content on ELISE dissemination activities to the large public and industries (M36) as part of the WP6 "Dissemination and communication".

Project website: www.elise-ai.eu

Diabetic Kidney Disease (DKD) is highly prevalent in type 2 diabetes, with major impact on patients and healthcare systems. The complex disorder, further modulated by cardiovascular comorbidities, presents as an accumulation of risk factors, which we treat with drug combinations. While the overall benefit of this approach is evident on a cohort level, individual patients show remarkable heterogeneity in drug response, and lack of guidance on personalized medication results in suboptimal control of the disorder. For resolving variability, we propose a new concept for personalization of drug combinations beyond the cohort-centric perspective. We improve patient stratification based on equivalence relations of clinical presentation, disease pathophysiology and drug combinations. The approach is derived from dynamical systems theory, aimed at reducing probabilistic assignment of patient-specific disease evolution and matching drug combinations. The availability of a large European repository holding DKD patients in routine care with diverse drug combinations, complemented by high-throughput screening for improving patient phenotyping, and molecular network modelling of pathology, embedded risk factor combinations and consequence of drug effect allows a systems representation of patient groups. Integrating clinical presentation and molecular architecture in a novel computational framework will establish a decision support software prototype. We will validate this tool for predicting optimized personalized drug combinations in a study using given clinical trial repositories. Demonstration will expand to other available drugs, which in combination with approved drugs promise benefit for groups of DKD patients. With a clear route toward uptake in the clinical setting, and generalization capacity of our approach to other complex disorders we foster next steps in personalization, anticipate major patient benefit, and see novel translation and business opportunities. 

Team of research: Michele Braccini, Veronica Distefano, Antonio Magnani, Maria Mannone, Claudio Silvestri, Debora Slanzi.

Budget for ECLT: 781,405.00 €
Project website: dc-ren.eu

Machine-learning technologies have become pervasive, and even able to outperform humans on specific tasks. However, it has been shown that they suffer from hallucinations known as adversarial examples, i.e., imperceptible, adversarial perturbations to images, text and audio that fool these systems into perceiving things that are not there. This has severely questioned their suitability for mission-critical applications, including self-driving cars and autonomous vehicles. The defense strategies proposed to overcome this issue have been shown to be ineffective against more sophisticated attacks carefully crafted to bypass them, highlighting the challenging nature of this problem. In this project, we formulate three main challenges that demand for novel learning paradigms, able to take reliable and explainable decisions, to assess and mitigate the security risks associated to such potential misuses of machine learning. This project will pave the way towards the design of reliable and explainable machines that are also useful beyond adversarial settings. We will indeed develop tools and prototypes that can face the challenges posed not only by cybersecurity applications with a clear adversarial nature, but also by recent computer-vision and deep-learning technologies.