SyReM shows that replaying memory samples whose loss gradients resemble the current batch, under a gradient-projection constraint, improves both stability and plasticity in online motion forecasting.
A comprehensive survey of forgetting in deep learning beyond continual learning,
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
unclear 1representative citing papers
citing papers explorer
-
Escaping Stability-Plasticity Dilemma in Online Continual Learning for Motion Forecasting via Synergetic Memory Rehearsal
SyReM shows that replaying memory samples whose loss gradients resemble the current batch, under a gradient-projection constraint, improves both stability and plasticity in online motion forecasting.