Reproduce “DRIVE: Distributional and Retrieval-Augmented Bidding with Value Evaluation” 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
Miduo Cui, Haochen Wang, Shangqin Mao, Xun Yang, Qianlong Xie, Xingxing Wang, Xuri Ge, Ying Zhou and 1 more
DRIVE addresses offline auto-bidding by generating multiple candidate bids and selecting among them with a value critic. Its Transformer models bid modes, while retrieval adds historical actions from similar contexts. On AuctionNet’s dense benchmark, DRIVE averages Value 386.6 across five budget settings, versus 378.4 for CQL; it does not lead at every budget, and Value does not establish CPA compliance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.