Researchers have developed TS-GEAR, a generative retrieval framework designed for on-demand delivery platforms like Taobao Shangou. This system addresses challenges such as geographic reachability, user preference variability across regions, and the need for set-level evaluation rather than individual sequence scoring. TS-GEAR incorporates Geo-Semantic ID for reachability prediction, Query-Geo Target Attention for personalized behavior aggregation, and Group Marginal Policy Optimization for improved candidate set diversity. When deployed on Taobao Shangou, TS-GEAR demonstrated superior performance in ad impressions and revenue compared to existing methods. AI
IMPACT This framework could improve the efficiency and revenue of on-demand delivery services by optimizing ad retrieval.
RANK_REASON The item is a research paper detailing a new framework and its performance evaluation. [lever_c_demoted from research: ic=1 ai=1.0]
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →