BCAM Researchers Present Their Work at NeurIPS 2025

• BCAM researchers Verónica Álvarez, Xabier de Juan, and Santiago Mazuelas attended the Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025) in San Diego (California).
• This is one of the most prestigious international conferences in machine learning and artificial intelligence.

BCAM researchers Verónica Álvarez, Xabier de Juan, and Santiago Mazuelas attended the Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS2025) in San Diego (California). This is one of the most prestigious international conferences in machine learning and artificial intelligence.

During the event, they presented three research papers showcasing their contributions to robust statistics, machine learning theory, and trustworthy AI:

• Santiago Mazuelas presented
  “Split Conformal Classification with Unsupervised Calibration”,
  which introduces new methods to enhance conformal prediction without requiring labeled data during calibration.
• Santiago Mazuelas and Xabier de Juan presented
  “On the Optimality of the Median-of-Means Estimator under Adversarial Contamination”,
  providing new insights into the robustness and optimality of classical estimators in adversarial settings.
• Santiago Mazuelas and Verónica Álvarez presented
  “Robust Minimax Boosting with Performance Guarantees”,
  a work that develops boosting techniques with minimax robustness and theoretical performance bounds.

Their participation highlights BCAM’s commitment to advancing fundamental and applied research in machine learning, statistical learning theory, and reliable AI systems.