Back to AI News
arxiv

TS-GEAR: Generative Ad Retrieval for On-Demand Delivery at Taobao Shangou

Researchers present TS-GEAR, a generative retrieval framework tailored to on-demand delivery platforms where merchants outside a user's delivery range are invalid candidates. The method introduces Geo-Semantic IDs that split merchant identifiers into geographic prefixes and semantic suffixes, Query-Geo Target Attention that conditions behavior aggregation on the live query and location, and Group Marginal Policy Optimization to reward collective set value. It is deployed in Taobao Shangou's search advertising pipeline with a 0.1B backbone, serving over 150 million queries per day. The authors report improvements in HitRate and geographic validity, alongside 3.8% gains in ad impressions and 4.2% in ad revenue.

AI Brief

It offers a concrete, production-deployed design for adapting generative retrieval to geography-constrained delivery advertising, with reported online gains.

arXiv:2611.12180v1 Announce Type: cross Abstract: On-demand delivery platforms such as DoorDash and Uber Eats impose a constraint absent in traditional e-commerce: merchants beyond the user's delivery range are invalid candidates rather than merely low-ranked. Generative retrieval has recently shown strong potential in large-scale retrieval systems. Adapting it to this setting, however, poses three challenges. First, existing semantic IDs encode geography only as a soft signal, so reachability can only rely on a non-differentiable post-generation hard truncation. Second, user preferences vary substantially across both regions and queries, so aggregating historical behaviors without conditioning on geography and the current query yields biased interest estimates. Third, the retrieved candidate set is evaluated by its collective value, yet autoregressive training scores each sequence independently, with no set-level signal to discourage redundant near-duplicates. To address the challenges, we propose TS-GEAR, a unified generative retrieval framework for on-demand delivery: (1) Geo-Semantic ID splits merchant identifiers into geographic prefixes and semantic suffixes, making reachability a prediction target rather than a post-hoc filter; (2) Query-Geo Target Attention aggregates historical behaviors using the real-time query and location as dual targets, while encoding each behavior's own location into its attention key; (3) GMPO (Group Marginal Policy Optimization) rewrites advantages using a deduplicated group reward and marginal advantage within GRPO, avoiding auxiliary diversity objectives. Deployed online with a lightweight 0.1B backbone, TS-GEAR serves over 150 million queries per day in the search advertising pipeline of Taobao Shangou, consistently outperforming state-of-the-art generative retrieval methods on both HitRate and geographic validity, and achieving 3.8% gains in ad impressions and 4.2% in ad revenue.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!