Research map / Reinforcement learning / Offline RL
Sustainable Online Reinforcement Learning for Auto-bidding
Zhiyu Mou, Yusen Huo, Rongquan Bai, Mingzhou Xie, Chuan Yu, Jian Xu, Bo Zheng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Train the bidding policy with V-CQL instead of USCB | Auto-bidding on TaoBao · TaoBao real-world A/B test, newly trained policy | BuyCnt: +3.20% | Improved |
| Train the bidding policy with V-CQL instead of USCB | Auto-bidding on TaoBao · TaoBao real-world A/B test, newly trained policy | ROI: +1.28% | Improved |
| Train the bidding policy with V-CQL instead of USCB | Auto-bidding on TaoBao · TaoBao real-world A/B test, newly trained policy | CPA: -2.03% | Improved |
| Train the bidding policy with V-CQL instead of USCB | Auto-bidding on TaoBao · TaoBao real-world A/B test, newly trained policy | ConBdg: +1.12% (newly trained USCB); -0.57% (safe policy) | Mixed |
| Shift the safe-policy sampling distribution toward the learned Q function (ablation) | SER exploration · TaoBao real-world A/B test versus safe policy | BuyCnt: -1.83% versus -3.32% | Improved |
| Shift the safe-policy sampling distribution toward the learned Q function (ablation) | SER exploration · TaoBao real-world A/B test versus safe policy | ConBdg: +0.67% versus +1.11% | Improved |
| Add the near-quadratic Q-function regularizer to CQL (ablation) | V-CQL · Simulated auto-bidding, 100 random seeds | performance variance | Improved |
| Use SER sampling rather than vanilla Gaussian sampling in the safety zone (ablation) | Online exploration · Simulated auto-bidding across iterations | BuyCnt | Improved |
| Guide SER exploration with Q functions from policies of different performance (ablation) | SER exploration · Simulated auto-bidding | BuyCnt: about 3.52% | Mixed |
Closest papers by meaning
- AuctionNet: A Novel Benchmark for Decision-Making in Large-Scale Games (NeurIPS 2024 · similarity 0.86)
- Learning Personalized Ad Impact via Contextual Reinforcement Learning under Delayed Rewards (NeurIPS 2025 · similarity 0.85)
- No-Regret Online Autobidding Algorithms in First-price Auctions (NeurIPS 2025 · similarity 0.85)
- Online Matching with Stochastic Rewards: Provable Better Bound via Adversarial Reinforcement Learning (ICML 2024 · similarity 0.85)
- LLM-Auction: Generative Auction towards LLM-Native Advertising (NeurIPS 2026 · similarity 0.84)
- A Perspective of Q-value Estimation on Offline-to-Online Reinforcement Learning (AAAI 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
