IRLLRec aligns LLM-generated textual intents with interaction-based intents via contrastive alignment and momentum distillation, improving recommendation accuracy on three datasets.
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
-
Intent Representation Learning with Large Language Model for Recommendation
IRLLRec aligns LLM-generated textual intents with interaction-based intents via contrastive alignment and momentum distillation, improving recommendation accuracy on three datasets.