GSTDS selects a scheduled fraction of each training batch via Fiedler-vector ranking and reference-model losses, reporting comparable or better accuracy on three small image benchmarks with up to 4x fewer training FLOPs, but with single-run results and an internal algorithm mismatch.
Journal of Machine Learning Research 19(27), 1–21 (2018)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Efficient Training of Deep Networks using Guided Spectral Data Selection: A Step Toward Learning What You Need
GSTDS selects a scheduled fraction of each training batch via Fiedler-vector ranking and reference-model losses, reporting comparable or better accuracy on three small image benchmarks with up to 4x fewer training FLOPs, but with single-run results and an internal algorithm mismatch.