AWhat changed
Reuses batched conjugate-gradient work to estimate solves, trace derivatives and log determinants using kernel matrix-matrix multiplication, paired with preconditioning and a PyTorch framework.
Scientific scope
Reuses batched conjugate-gradient work to estimate solves, trace derivatives and log determinants using kernel matrix-matrix multiplication, paired with preconditioning and a PyTorch framework. Finite iterative solves, stochastic probes and partial Lanczos quadrature retain numerical/sampling error; a low-rank preconditioner is not necessarily a low-rank approximation to the GP model. One solver call still contains many iterations, not one matrix multiply. Gardner/Pleiss equal contributors; CG/Lanczos/trace-estimation foundations inherited.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Finite iterative solves, stochastic probes and partial Lanczos quadrature retain numerical/sampling error; a low-rank preconditioner is not necessarily a low-rank approximation to the GP model. One solver call still contains many iterations, not one matrix multiply. Gardner/Pleiss equal contributors; CG/Lanczos/trace-estimation foundations inherited.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction