Adversarial Adaptive Framework for Cold-Start Cross-Domain Recommendation
This article presents an adversarial adaptive framework that aligns source and target domains to address domain shift and severe data imbalance in cold-start cross-domain recommendation, demonstrating significant CTR and CVR performance gains when combined with various state‑of‑the‑art single‑domain models.