A gradient-alignment influence score (GGscore) that selects the highest- and lowest-scoring old interactions for replay improves incremental neural recommendation slightly over random replay, mainly at large replay ratios.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MEGG: Replay via Maximally Extreme GGscore in Incremental Learning for Neural Recommendation Models
A gradient-alignment influence score (GGscore) that selects the highest- and lowest-scoring old interactions for replay improves incremental neural recommendation slightly over random replay, mainly at large replay ratios.