Trajectory-wise Iterative Reinforcement Learning Framework for Auto-bidding

arXiv:2402.15102v1 Announce Type: new
Abstract: In online advertising, advertisers participate in ad auctions to acquire ad opportunities, often by utilizing auto-bidding tools provided by demand-side platforms (DSPs). The current auto-bidding algorithms typically employ reinforcement learning (RL). However, due to safety concerns, most RL-based auto-bidding policies are trained in simulation, leading to a performance degradation when deployed in online environments. To narrow this gap, we can deploy multiple auto-bidding agents in parallel to collect a large interaction dataset. Offline RL algorithms …

abstract advertisers advertising algorithms arxiv auto bidding concerns cs.ai cs.gt cs.ir cs.lg current demand framework iterative online advertising opportunities performance platforms reinforcement reinforcement learning safety simulation tools trajectory type wise