Reproduce “Deep Automated Mechanism Design for Integrating Ad Auction and Allocation in Feed” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xuejian Li, Ze Wang, Bingqi Zhu, Fei He, Yongkang Wang, Xingxing Wang
Researchers at Meituan developed MIAA, a deep automated mechanism that integrates ad auction and allocation by explicitly modeling global externalities and ensuring incentive compatibility. This integrated approach led to an 8.17% improvement in 'Revenue + αGMV' during online A/B tests on Meituan's retail delivery platform.
Key insights
- Explicitly modeling global externalities (position and context) within ad prediction significantly improves CTR prediction accuracy.
- Integrating ad auction and allocation decisions, while maintaining economic properties like Incentive Compatibility, yields superior platform performance compared to traditional separated approaches.
- Deep learning can automate the design and optimization of complex auction mechanisms, making them adaptable and efficient for real-world deployment.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.