# An end-to-end Generative Retrieval Method for Sponsored Search Engine --Decoding Efficiently into a Closed Target Domain

arXiv: https://arxiv.org/abs/1902.00592 · HTML: https://mlsanity.com/paper/1902.00592/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Feb 2019.

## Built on by (most cited first, 12 of 12)

- [DeepXML: A Deep Extreme Multi-Label Learning Framework Applied to Short Text Documents](https://mlsanity.com/paper/2111.06685/) (2021)
- [Diversity driven Query Rewriting in Search Advertising](https://mlsanity.com/paper/2106.03816/) (2021)
- [Triangular Bidword Generation for Sponsored Search Auction](https://mlsanity.com/paper/2101.11349/) (2021)
- [Scaling the Vocabulary of Non-autoregressive Models for Efficient Generative Retrieval](https://mlsanity.com/paper/2406.06739/) (2024)
- [Improving Retrieval in Sponsored Search by Leveraging Query Context Signals](https://mlsanity.com/paper/2407.14346/) (2024)
- [Modified Query Expansion Through Generative Adversarial Networks for Information Extraction in E-Commerce](https://mlsanity.com/paper/2301.00036/) (2022)
- [CPRM: A LLM-based Continual Pre-training Framework for Relevance Modeling in Commercial Search](https://mlsanity.com/paper/2412.01269/) (2024)
- [OKG: On-the-Fly Keyword Generation in Sponsored Search Advertising](https://mlsanity.com/paper/2412.03577/) (2024)
- [Multi-objective Aligned Bidword Generation Model for E-commerce Search Advertising](https://mlsanity.com/paper/2506.03827/) (2025)
- [OMS: On-the-fly, Multi-Objective, Self-Reflective Ad Keyword Generation via LLM Agent](https://mlsanity.com/paper/2507.02353/) (2025)
- [Modular Deep Learning for Multivariate Time-Series: Decoupling Imputation and Downstream Tasks](https://mlsanity.com/paper/2411.03941/) (2024)
- [One Model, Two Markets: Bid-Aware Generative Recommendation](https://mlsanity.com/paper/2603.22231/) (2026)

## Abstract (arXiv)

In this paper, we present a generative retrieval method for sponsored search engine, which uses neural machine translation (NMT) to generate keywords directly from query. This method is completely end-to-end, which skips query rewriting and relevance judging phases in traditional retrieval systems. Different from standard machine translation, the target space in the retrieval setting is a constrained closed set, where only committed keywords should be generated. We present a Trie-based pruning technique in beam search to address this problem. The biggest challenge in deploying this method into a real industrial environment is the latency impact of running the decoder. Self-normalized training coupled with Trie-based dynamic pruning dramatically reduces the inference time, yielding a speedup of more than 20 times. We also devise an mixed online-offline serving architecture to reduce the latency and CPU consumption. To encourage the NMT to generate new keywords uncovered by the existing system, training data is carefully selected. This model has been successfully applied in Baidu's commercial search engine as a supplementary retrieval branch, which has brought a remarkable revenue improvement of more than 10 percents.