AWhat changed
Trains a stochastic supernetwork using a Gumbel-softmax relaxation and latency-aware objective to select layer-wise blocks for mobile inference.
Scientific scope
Trains a stochastic supernetwork using a Gumbel-softmax relaxation and latency-aware objective to select layer-wise blocks for mobile inference. Differentiable search selected device-specific network architectures; researchers designed the framework, measurements and validation. Latency is estimated by summing separately benchmarked operators, assuming near-independent sequential runtimes; the proxy need not hold on all hardware. Search-space macroarchitecture/filter choices remain hand designed. First-author star denotes Facebook internship, not equal contribution.
Evidence
Differentiable search selected device-specific network architectures; researchers designed the framework, measurements and validation. WRS has not independently rerun these experiments. Differentiable search selected device-specific network architectures; researchers designed the framework, measurements and validation. Latency is estimated by summing separately benchmarked operators, assuming near-independent sequential runtimes; the proxy need not hold on all hardware. Search-space macroarchitecture/filter choices remain hand designed. First-author star denotes Facebook internship, not equal contribution.
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