AWhat changed
Matched score differences along random directions and used integration by parts to obtain objectives computable with Hessian-vector products, including a variance-reduced form for scalable score estimation.
Scientific scope
Matched score differences along random directions and used integration by parts to obtain objectives computable with Hessian-vector products, including a variance-reduced form for scalable score estimation. Projection covariance/moment and boundary regularity assumptions are required. Consistency results additionally assume a well-specified identifiable model and other conditions; they do not guarantee finite neural optimization success. A directly fitted vector field need not be an integrable scalar-density gradient.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Projection covariance/moment and boundary regularity assumptions are required. Consistency results additionally assume a well-specified identifiable model and other conditions; they do not guarantee finite neural optimization success. A directly fitted vector field need not be an integrable scalar-density gradient.
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