Researchers from across the Valencian Region present projects on the application of AI in healthcare, translation, profile identification, and understanding large language models
Valencia, July 25, 2024.- The Universitat Politècnica de València hosted the 4th JRC (Joint Research Unit) Meeting, which brought together prominent researchers and AI experts from various universities and research centers to share the latest advances in this field and promote interdisciplinary collaboration.
Cesar Ferri from the Universitat Politècnica de València explored in his presentation how machine learning algorithms can justify their decisions in critical contexts. He highlighted the use of teaching algorithms to identify a small set of labeled examples, allowing users to build their own model of the AI system and better understand the target concept. The research covered domains such as Boolean functions and time series, offering new perspectives on explaining automated decisions in AI.
Vicente R. Tomás from Universitat Jaume I introduced “An intelligent transport system for safe route recommendation”. This innovative system uses intelligent agents to manage traffic in a distributed and flexible way, providing safe route recommendations using a statistical model based on real data and deep learning. This advanced technology not only analyses current traffic conditions, but also predicts possible incidents, thus optimising routes and improving road safety.
Amparo Alcina from the Universitat Jaume I discussed her work on the ontological modeling of language and the representation of linguistic knowledge. Her research stood out for applying computing tools and techniques in professional translation and localization, as well as in terminology and translation studies.
For his part, José Mª Azorín from the Universidad Miguel Hernández de Elche presented his innovative project on Brain-Machine Interfaces (BMI) for gait assistance with lower limb exoskeletons. This work, aimed at improving the rehabilitation of people with motor disabilities due to strokes and spinal cord injuries, demonstrated how brain-machine interfaces can control exoskeletons more intuitively and effectively, enhancing neuroplasticity and offering new opportunities in the field of rehabilitation.
Additionally, Irene Epifanio from the Universitat Jaume I highlighted her advances in archetypal analysis and its application in various fields, including industrial design and human development. Her research underscores how archetypal analysis can identify extreme profiles and improve the understanding of multivariate and functional data, providing a valuable tool for various scientific and social applications.
The last speaker, Juan Carlos Trujillo from Universitat d’Alacant, addressed the understanding of large language models through mechanical interpretation (MI). This is achieved through the analysis of multiple consecutive tokens using MI, laying the foundations for understanding more complex behaviors, increasing the safety and reliability of an AI system.
Participants had the opportunity to discuss and explore potential synergies between their projects, thereby strengthening the AI research network in the Valencian Region. The success of the meeting reaffirms the fundamental role of the ValgrAI UMI as a catalyst for innovation and scientific advancement in the Valencian Region.
The 4th UMI Meeting not only stood out for the quality and relevance of the presented projects but also for the emphasis on AI applications in healthcare, which opens new avenues for improving patients’ quality of life through advanced technologies. AI research continues to evolve, and events like this are essential to ensure that scientific developments have a positive and tangible impact on society.
