Compartir

En els últims anys, l’ús de proxies per a desmitificar algorismes basats en caixes negres ha adquirit popularitat, fins a convertir-se en metodologies estàndard en entorns industrials. No obstant això, encara existeixen desafiaments quant a la interpretació dels resultats i la generalització d’aquestes explicacions en escenaris crítics, on la precisió i la transparència són primordials.

Programa

  • Introducció a Explainable AI (XAI).
  • Limitació de les tècniques XAI actuals en termes d’interpretabilitat
  • Característiques d’explicacions usables entorns crítics.
  • Tendències a futur de XAI.

Masterclass impartida per la Fundació i2Cat

Impartit per:

ALBERT CALVO
Senior AI & Cyber Researcher, Fundació i2CAT

Albert Calvo works as an AI Researcher at the i2cat foundation in the Department of Distributed Artificial Intelligence (DAI) and also is an associate professor in the Department of Computer Science at the UPC (Polytechnic University of Catalonia). Albert is a computer engineer specializing in Information Technologies. He also pursued a Master’s degree in Computer Science, where he conducted his master’s thesis at EPFL (École polytechnique fédérale de Lausanne). Currently, Albert is a doctoral candidate in computer science at the UPC. His research focuses on Explainability in industrial projects related to Data Science, aiming to build solutions that meet the requirements of transparency and robustness, allowing them to be used seamlessly by the end user. The researcher has also authored publications in internationally renowned conferences and journals in the field of Artificial Intelligence, including ECAI, DSAA, and Data Mining and Knowledge Discovery, as well as presentations at cybersecurity conferences such as FIRST and eCrime Symposium.

CIDAI