SparseRec prunes recommendation embedding tables during training using NMF-based initialization, cumulative gradients for regrowth, and sparse backward passes, beating baselines on two datasets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Sparser Training for On-Device Recommendation Systems
SparseRec prunes recommendation embedding tables during training using NMF-based initialization, cumulative gradients for regrowth, and sparse backward passes, beating baselines on two datasets.