![[2602.22226] SEGB: Self-Evolved Generative Bidding with Local Autoregressive Diffusion](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22226] SEGB: Self-Evolved Generative Bidding with Local Autoregressive Diffusion
Summary
The paper presents Self-Evolved Generative Bidding (SEGB), a novel framework for automated online advertising that enhances bidding strategies by synthesizing future states and refining policies without external input.
Why It Matters
SEGB addresses the limitations of existing generative bidding methods by enabling dynamic foresight and self-improvement from static data. This innovation is crucial for advertisers seeking to optimize their bidding strategies in real-time, thereby increasing efficiency and business value.
Key Takeaways
- SEGB synthesizes plausible future states to enhance bidding decisions.
- The framework allows for self-guided policy refinement without external intervention.
- Experiments show SEGB significantly outperforms existing bidding strategies.
- The approach delivered a +10.19% increase in target cost during real-world applications.
- SEGB's self-contained method enables robust policy improvement from static data.
Computer Science > Information Retrieval arXiv:2602.22226 (cs) [Submitted on 31 Dec 2025] Title:SEGB: Self-Evolved Generative Bidding with Local Autoregressive Diffusion Authors:Yulong Gao, Wan Jiang, Mingzhe Cao, Xuepu Wang, Zeyu Pan, Haonan Yang, Ye Liu, Xin Yang View a PDF of the paper titled SEGB: Self-Evolved Generative Bidding with Local Autoregressive Diffusion, by Yulong Gao and 7 other authors View PDF HTML (experimental) Abstract:In the realm of online advertising, automated bidding has become a pivotal tool, enabling advertisers to efficiently capture impression opportunities in real-time. Recently, generative auto-bidding has shown significant promise, offering innovative solutions for effective ad optimization. However, existing offline-trained generative policies lack the near-term foresight required for dynamic markets and usually depend on simulators or external experts for post-training improvement. To overcome these critical limitations, we propose Self-Evolved Generative Bidding (SEGB), a framework that plans proactively and refines itself entirely offline. SEGB first synthesizes plausible short-horizon future states to guide each bid, providing the agent with crucial, dynamic foresight. Crucially, it then performs value-guided policy refinement to iteratively discover superior strategies without any external intervention. This self-contained approach uniquely enables robust policy improvement from static data alone. Experiments on the AuctionNet benchma...