Researchers have developed a new framework called TWICE to improve long-horizon conversion prediction in online advertising. This method addresses the challenge of delayed feedback by separating conversion probability from the delay distribution. TWICE was tested in an A/B test on Kwai's advertising system, where it led to significant increases in expected revenue, revenue, and conversions, prompting its full deployment. AI
IMPACT This framework could enhance the effectiveness of online advertising systems by improving prediction accuracy for long-term conversion rates.
RANK_REASON The cluster describes a new research paper detailing a novel framework for conversion prediction, including experimental results and deployment in a real-world system. [lever_c_demoted from research: ic=1 ai=0.7]
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
