BWhat changed
Combined factorization-machine interaction blocks, learned low-rank projections and linear compression blocks in a residual stack, with coordinated depth/width scaling for large recommendation datasets.
Scientific scope
Combined factorization-machine interaction blocks, learned low-rank projections and linear compression blocks in a residual stack, with coordinated depth/width scaling for large recommendation datasets. Empirical scaling over the tested compute/model/data ranges is not an all-scale law. Low-rank interaction compression depends on dimensions and learned projections; the interaction-order explanation does not certify learning every interaction. Distinct from DHEN's heterogeneous-module ensemble; inherited factorization machines and residual normalization are not new here.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Empirical scaling over the tested compute/model/data ranges is not an all-scale law. Low-rank interaction compression depends on dimensions and learned projections; the interaction-order explanation does not certify learning every interaction. Distinct from DHEN's heterogeneous-module ensemble; inherited factorization machines and residual normalization are not new here.
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