A new system called CASH has been developed to optimize advertiser subsidy allocation on advertising platforms. CASH utilizes a combination of a knowledge-guided large language model and a data-driven numerical response model to learn from past outcomes and allocate subsidies at an individual account level. An online experiment showed that CASH increased revenue per account by 15.35% and ADVV by 19.86% compared to a control group, while maintaining a similar subsidy rate. AI
IMPACT This system demonstrates how LLMs can be integrated with numerical models to optimize business processes, potentially improving efficiency and revenue in advertising platforms.
RANK_REASON The item is a research paper published on arXiv detailing a new system for advertiser subsidy allocation. [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 →