Snapshots of Deep Learning Model Error Bounds in Credit Derivatives Valuation

Quickly apply original, key PMR-published papers with Snapshots—a short article companion that distills PMR research into compressed, digestible takeaways, so you can put the paper’s core ideas to work in your investment process—fast. This Snapshot article is based on research arguing that deep learning models approximated CDS index and CDS option prices with generally sub-1% errors and large speedups, while differential machine learning supported intuitive gradient-based error bars that captured most tested ground-truth prices.