# SMART: LLM-Augmented Hybrid Retrieval for Dynamic Product Ads

arXiv: https://arxiv.org/abs/2607.23121 · HTML: https://mlsanity.com/paper/2607.23121/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 25 references that are in the map.
- First posted on arXiv 25 Jul 2026.

## Builds on (most cited first, 13 of 13)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection](https://mlsanity.com/paper/2310.11511/) (2023)
- [SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features](https://mlsanity.com/paper/2502.14786/) (2025)
- [Precise Zero-Shot Dense Retrieval without Relevance Labels](https://mlsanity.com/paper/2212.10496/) (2022)
- [Hybrid LLM: Cost-Efficient and Quality-Aware Query Routing](https://mlsanity.com/paper/2404.14618/) (2024)
- [Large Language Models for Information Retrieval: A Survey](https://mlsanity.com/paper/2308.07107/) (2023)
- [Sparse, Dense, and Attentional Representations for Text Retrieval](https://mlsanity.com/paper/2005.00181/) (2020)
- [Recommender Systems with Generative Retrieval](https://mlsanity.com/paper/2305.05065/) (2023)
- [Query2doc: Query Expansion with Large Language Models](https://mlsanity.com/paper/2303.07678/) (2023)
- [AutoMix: Automatically Mixing Language Models](https://mlsanity.com/paper/2310.12963/) (2023)
- [LLM-based Query Expansion Fails for Unfamiliar and Ambiguous Queries](https://mlsanity.com/paper/2505.12694/) (2025)
- [COEF-VQ: Cost-Efficient Video Quality Understanding through a Cascaded Multimodal LLM Framework](https://mlsanity.com/paper/2412.10435/) (2024)
- [AdNanny: One Reasoning LLM for All Offline Ads Recommendation Tasks](https://mlsanity.com/paper/2602.01563/) (2026)

## Abstract (arXiv)

Dynamic Product Ads (DPA) require retrieving relevant items from multi-million product catalogs, balancing two competing objectives: retargeting (re-surfacing known interests) and prospecting (discovering new categories). While Large Language Models (LLMs) capture semantic intent better than traditional embedding models, deploying them at scale introduces prohibitive inference costs and lexical mismatch issues. Through controlled experiments on millions of users, we demonstrate a critical retrieval decomposition: rule-generated queries excel at retargeting on a lexical BM25 index, while LLM-generated queries excel at prospecting on a dense ANN index. Building on this, we propose SMART (SeMantic-aware Adaptive ReTrieval). To manage costs, a lightweight quality gate identifies coverage gaps in initial keyword results, adaptively routing only the ~10% of users who benefit from semantic prospecting to the LLM path. Offline evaluation demonstrates that this gated approach captures the bulk of semantic prospecting gains in Relevance Score while maintaining competitive re-targeting performance at a 90% reduction in LLM costs. Finally, in a 2-week online A/B test at Snap, SMART improved the ad conversion rate by +27.6% over a strong embedding-based baseline.