's likelihood kernel that bounds the loss of information based on information-geometric

Learning curve extrapolation aims to predict model performance in later epochs of training, based on the performance in earlier epochs. In this work, we argue that, while the inherent uncertainty in the extrapolation of learning curves warrants a Bayesian approach, existing methods are (i) overly restrictive, and/or (ii) computationally expensive.

In this section, we first provide a summary of the notation in Table A2.