Dual-LS pairs reservoir and gradient-diversity replay buffers with fast and slow EMA model updates to mitigate catastrophic forgetting in task-free online vehicle motion forecasting.
Title resolution pending
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
-
Complementary Learning System Empowers Online Continual Learning of Vehicle Motion Forecasting in Smart Cities
Dual-LS pairs reservoir and gradient-diversity replay buffers with fast and slow EMA model updates to mitigate catastrophic forgetting in task-free online vehicle motion forecasting.