AWhat changed
Clustered training queries once, sampled topically related batches and balanced positive/negative passage pairs by teacher-score margins, combining pairwise cross-encoder and in-batch ColBERT teacher signals.
Scientific scope
Clustered training queries once, sampled topically related batches and balanced positive/negative passage pairs by teacher-score margins, combining pairwise cross-encoder and in-batch ColBERT teacher signals. Teacher inference, reused teacher scores and one-time query clustering have costs even when omitted from the student loop. Single-GPU timing and retrieval latency use the stated hardware/index configuration. Balanced teacher margins do not certify relevance labels or eliminate false negatives.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Teacher inference, reused teacher scores and one-time query clustering have costs even when omitted from the student loop. Single-GPU timing and retrieval latency use the stated hardware/index configuration. Balanced teacher margins do not certify relevance labels or eliminate false negatives.
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