ViRN combines variational inference with Wasserstein-distance neighbor fusion to estimate tail-class distributions in continual learning, reporting gains on six long-tailed benchmarks.
A unified continual learning framework with general parameter-efficient tuning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ViRN: Variational Inference and Distribution Trilateration for Long-Tailed Continual Representation Learning
ViRN combines variational inference with Wasserstein-distance neighbor fusion to estimate tail-class distributions in continual learning, reporting gains on six long-tailed benchmarks.