AWhat changed
Trained multiple nested prefixes of a shared embedding with size-specific objectives, enabling downstream classification/retrieval to use different dimensionalities without separate encoder forward passes.
Scientific scope
Trained multiple nested prefixes of a shared embedding with size-specific objectives, enabling downstream classification/retrieval to use different dimensionalities without separate encoder forward passes. Multi-prefix loss adds training obligations; the encoder forward cost is not reduced merely by truncating its output. Quality preservation, interpolated dimensions and speedups are empirical for tested setups, not every dimension/task. Nested-dropout and other adaptive-network predecessors are acknowledged; first3equal credit includes a project-lead distinction.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Multi-prefix loss adds training obligations; the encoder forward cost is not reduced merely by truncating its output. Quality preservation, interpolated dimensions and speedups are empirical for tested setups, not every dimension/task. Nested-dropout and other adaptive-network predecessors are acknowledged; first3equal credit includes a project-lead distinction.
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