Unified Supervision for Sponsored Search Retrieval CCS: Applied computing E-commerce infrastructure CCS: Information systems Novelty in information retrieval
Shasvat Desai, Md Omar Faruk Rokon, Jhalak Nilesh Acharya, Isha Shah, Hong Yao, Utkarsh Porwal, Kuang-chih Lee · 2026
Read Unified Supervision for Sponsored Search Retrieval CCS: Applied computing E-commerce infrastructure CCS: Information systems Novelty in information retrieval by Shasvat Desai, Md Omar Faruk Rokon, Jhalak Nilesh Acharya, Isha Shah, Hong Yao, Utkarsh Porwal, Kuang-chih Lee (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-07930.txt.
Description
Desai and six colleagues improve Walmart sponsored search retrieval by unifying training signals. Fast first-stage retrievers usually learn from user clicks, logged at scale without annotation. Clicks are a flawed proxy for relevance: they may reflect popularity, promotion, visuals, titles, or price rather than query fit. Sponsored search worsens the problem, since ad exposure depends on auctions and advertiser moves, leaving click data sparse by design. Their unified training blends semantic relevance with click signals so retrievers rank items users truly want, not merely items they happened to see.
Catalog record
| CREATOR | Shasvat Desai, Md Omar Faruk Rokon, Jhalak Nilesh Acharya, Isha Shah, Hong Yao, Utkarsh Porwal, Kuang-chih Lee |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 16362 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.07930