Full Professor, ELLIS Institute Finland
3 papers at NeurIPS 2025
We introduce a novel framework that seamlessly integrates amortized Bayesian inference and active data acquisition, featuring adaptive strategies that can optimize for diverse, user-specified learning objectives at deployment.
RCaGP is a general-purpose Gaussian Process framework that combines robustness to outliers and approximation-aware uncertainty—two coupled limitations of sparse GPs that are critical to tackle for challenging tasks like high-throughput BayesOpt.
This paper investigates a principle and develops a method for intervening on a single, targeted agent in a multi-agent reinforcement learning to reach a collective objective consisting of the primary task goal and an additional desired outcome.