The ValgrAI Foundation held the VII Meeting of the ValgrAI Joint Research Unit (UMI), which was part of the IV edition of the ValgrAI Scientific Council Forum 2025. The UMI brings together researchers in Artificial Intelligence of the Valencian Community, with the highest scientific production and successful cases of technology transfer in the field of AI and associated disciplines.
The event was opened by the General Director of ValgrAI, Vicent Botti, and the Managing Director of ValgrAI, Ana Cidad, who highlighted the importance of promoting collaborative, interdisciplinary research committed to social welfare, as well as the strategic role of ValgrAI as a platform to promote talent, responsible innovation and knowledge transfer in the field of artificial intelligence.
The first speaker was Juan Antonio Pérez Ortiz, from the Universitat d’Alacant, who presented the progress of the Transducens group in improving machine translation for digitally underrepresented languages, an initiative that seeks to democratize access to language technologies. The project is investigating new strategies such as multilingual translation agents, assisted reasoning and learning from grammar books, all aimed at ensuring that next-generation language models effectively incorporate the thousands of minority languages currently ignored by mainstream AI systems.
The session continued with the intervention of Rafael Berlanga, from the Universitat Jaume I de Castelló (UJI), who presented the results of the “Towards AI-empowered Cyber-SOCs” project, focused on improving computer security through AI tools applied to cybersecurity operations centers (CyberSOCs). Berlanga explained how, through deep learning models and automatic reasoning techniques, real threats can be detected more accurately among the large volume of alerts generated by systems, thus helping to reduce analyst fatigue and increase the effectiveness of incident response.
Joaquín Torres-Sospedra (Universitat de València) presented the latest advances of the Urban Computing Lab in indoor positioning using machine learning. His intervention addressed projects such as UJIIndoorLoc and new applications based on BLE sensors, WiFi and advanced cameras, oriented to industrial and mobility environments, demonstrating the potential of AI to offer robust solutions in complex contexts.
Jordi Linares-Pellicer, researcher at VRAIN-UPV, who addressed the concept of AI-mediated interfaces and how they are radically transforming the way people interact with technology. Linares presented a scenario in which traditional systems give way to interfaces generated in real time by language models, capable of adapting to the needs, capabilities and context of each user. This trend represents a revolution in both accessibility and user experience, anticipating the end of rigid interfaces and betting on a truly ubiquitous, transparent and personalized AI.
One of the most emotional moments of the meeting was the presentation of the project “Salvem les fotos”, focused on the recovery of family images affected by the DANA that hit the Valencian Community in 2024. José Daniel Galván and Hugo Albert (UPV) showed how AI can be put at the service of emotional memory, using neural networks to digitally segment and restore damaged photographs. This work is part of a collaborative initiative between public universities and other cultural and social entities, and has so far allowed the recovery of more than 200,000 images.
This VII UMI Meeting reaffirms ValgrAI’s commitment to people-centered artificial intelligence, driving shared knowledge and generating real impact in key sectors such as cybersecurity, linguistic inclusion, culture or intelligent mobility. In the words of the organizers, “the future of AI is not only about technological development, but also about its ability to respond in an ethical, critical and useful way to the great social and human challenges of the present”.
