AWhat changed
Ranked trained convolutional filters by absolute weight sums and removed corresponding feature maps and downstream kernels, producing smaller dense operations rather than isolated sparse weights.
Scientific scope
Ranked trained convolutional filters by absolute weight sums and removed corresponding feature maps and downstream kernels, producing smaller dense operations rather than isolated sparse weights. Filter magnitude is a practical proxy, with layer sensitivity and pruning ratios empirically selected. Greedy multi-layer selection is not globally optimal and residual links impose structural constraints. Accuracy recovery and speed depend on retraining, architecture and implementation, not only nominal filter/FLOP reductions.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Filter magnitude is a practical proxy, with layer sensitivity and pruning ratios empirically selected. Greedy multi-layer selection is not globally optimal and residual links impose structural constraints. Accuracy recovery and speed depend on retraining, architecture and implementation, not only nominal filter/FLOP reductions.
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