AWhat changed
Used residual gated convolutions for source and target sequence representations, with separate attention at decoder layers and parallel computation across training positions.
Scientific scope
Used residual gated convolutions for source and target sequence representations, with separate attention at decoder layers and parallel computation across training positions. Training parallelism does not remove autoregressive dependencies during generation. GLUs, residual learning, convolutions and attention are inherited components. Benchmarked speed/quality improvements are not a universal advantage over recurrent or Transformer models.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Training parallelism does not remove autoregressive dependencies during generation. GLUs, residual learning, convolutions and attention are inherited components. Benchmarked speed/quality improvements are not a universal advantage over recurrent or Transformer models.
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