A framework for real-time LLM-based user interest personas in large-scale video recommendations, using distillation, async inference, and video clustering to balance interests with novel topics and improve viewer value via A/B tests.
InProceedings of the Nineteenth ACM Conference on Recommender Systems (RecSys ’25)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Post-stratification plus CUPED cuts required traffic by about 45% for reliable A/B tests on heavy-tailed revenue metrics in ranking experiments.
citing papers explorer
-
LLM-Based User Personas for Recommendations at Scale
A framework for real-time LLM-based user interest personas in large-scale video recommendations, using distillation, async inference, and video clustering to balance interests with novel topics and improve viewer value via A/B tests.