IDAL combines ResNet-50 with FPN and a new pseudo-label-weighted MMD loss to improve unsupervised domain adaptation accuracy on natural image benchmarks, but the gains are small and inconsistent across datasets.
f-domain adver- sarial learning: Theory and algorithms
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
IDAL: Improved Domain Adaptive Learning for Natural Images Dataset
IDAL combines ResNet-50 with FPN and a new pseudo-label-weighted MMD loss to improve unsupervised domain adaptation accuracy on natural image benchmarks, but the gains are small and inconsistent across datasets.