For neural-network reparameterized full waveform inversion, incorporating the initial model by denormalization (direct addition) outperforms pretraining in accuracy, speed, and workflow simplicity on Marmousi experiments.
A survey on negative transfer,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
Initial Model Incorporation for Deep Learning FWI: Pretraining or Denormalization?
For neural-network reparameterized full waveform inversion, incorporating the initial model by denormalization (direct addition) outperforms pretraining in accuracy, speed, and workflow simplicity on Marmousi experiments.