Researchers have developed STAR, a framework for post-click conversion rate (PCVR) prediction in recommender systems. This framework addresses challenges with heterogeneous features, user sequences, and target-aware interests. STAR integrates structured feature tokenization with target-aware interest representation, building upon a HyFormer-style backbone. Key innovations include high-cardinality signal recovery, explicit user-item interaction tokens, and a contrastive auxiliary objective inspired by InfoNCE. The system also aligns training and inference pipelines to ensure robustness. AI
IMPACT This framework could improve the accuracy and robustness of recommender systems, leading to better user experiences and more effective advertising.
RANK_REASON The cluster contains an academic paper detailing a new framework for a specific machine learning task.
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
