C2:Cross learning module enhanced decision transformer with Constraint-aware loss for auto-bidding

TL;DR

This work addresses two main limitations of Decision Transformer in auto-bidding: weak cross-sequence correlations among state, action, and RTG, and indiscriminate learning of historical bidding patterns. It introduces Cross Learning Block (CLB) to strengthen inter-sequence modeling via cross-attention and Constraint-aware Loss (CL) to bias learning toward constraint-compliant, high-value trajectories. Evaluated on AuctionNet, C2 yields up to 3.23% gains over the prior SOTA GAVE, with ablations confirming the additive and synergistic benefits of CLB and CL. The framework demonstrates improved cross-correlation capture and constrained learning, with code released for reproducibility.

Abstract

Decision Transformer (DT) shows promise for generative auto-bidding by capturing temporal dependencies, but suffers from two critical limitations: insufficient cross-correlation modeling among state, action, and return-to-go (RTG) sequences, and indiscriminate learning of optimal/suboptimal behaviors. To address these, we propose C2, a novel framework enhancing DT with two core innovations: (1) a Cross Learning Block (CLB) via cross-attention to strengthen inter-sequence correlation modeling; (2) a Constraint-aware Loss (CL) incorporating budget and Cost-Per-Acquisition (CPA) constraints for selective learning of optimal trajectories. Extensive offline evaluations on the AuctionNet dataset demonstrate consistent performance gains (up to 3.2% over state-of-the-art method) across diverse budget settings; ablation studies verify the complementary synergy of CLB and CL, confirming C2's superiority in auto-bidding. The code for reproducing our results is available at: https://github.com/Dingjinren/C2.

Figures (2)

• Figure 1: Overall structure of C2.
• Figure 2: Cross-correlation learning ability: C2 vs. DT