A new causal disentanglement objective for recommendation models improves out-of-distribution generalization under policy-induced distribution shift, with production A/B gains despite offline parity.
Utilizing non-click samples via semi-supervised learning for conversion rate prediction
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
PRISM improves e-commerce search robustness by modeling preference-relevance interactions via preference rectification, LLM-driven semantic anchoring with prototypes, and preference-conditioned evidence routing.
citing papers explorer
-
Causal Representation Learning for Generalisable Recommendation
A new causal disentanglement objective for recommendation models improves out-of-distribution generalization under policy-induced distribution shift, with production A/B gains despite offline parity.
-
PRISM: Refracting the Entangled User Behavior Space for E-Commerce Search
PRISM improves e-commerce search robustness by modeling preference-relevance interactions via preference rectification, LLM-driven semantic anchoring with prototypes, and preference-conditioned evidence routing.