# Multimodal Generative Retrieval Model with Staged Pretraining for Food Delivery on Meituan

arXiv: https://arxiv.org/abs/2602.06654 · HTML: https://mlsanity.com/paper/2602.06654/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 37 references that are in the map.
- First posted on arXiv 6 Feb 2026.

## Builds on (most cited first, 19 of 19)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT](https://mlsanity.com/paper/2004.12832/) (2020)
- [DKN: Deep Knowledge-Aware Network for News Recommendation](https://mlsanity.com/paper/1801.08284/) (2018)
- [Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese](https://mlsanity.com/paper/2211.01335/) (2022)
- [Towards Understanding the Overfitting Phenomenon of Deep Click-Through Rate Prediction Models](https://mlsanity.com/paper/2209.06053/) (2022)
- [QARM: Quantitative Alignment Multi-Modal Recommendation at Kuaishou](https://mlsanity.com/paper/2411.11739/) (2024)
- [FORGE: Forming Semantic Identifiers for Generative Retrieval in Industrial Datasets](https://mlsanity.com/paper/2509.20904/) (2025)
- [Generative Dense Retrieval: Memory Can Be a Burden](https://mlsanity.com/paper/2401.10487/) (2024)
- [MMQ: Multimodal Mixture-of-Quantization Tokenization for Semantic ID Generation and User Behavioral Adaptation](https://mlsanity.com/paper/2508.15281/) (2025)
- [DAS: Dual-Aligned Semantic IDs Empowered Industrial Recommender System](https://mlsanity.com/paper/2508.10584/) (2025)
- [Hi-Gen: Generative Retrieval For Large-Scale Personalized E-commerce Search](https://mlsanity.com/paper/2404.15675/) (2024)
- [LEMUR: Large scale End-to-end MUltimodal Recommendation](https://mlsanity.com/paper/2511.10962/) (2025)
- [Progressive Semantic Residual Quantization for Multimodal-Joint Interest Modeling in Music Recommendation](https://mlsanity.com/paper/2508.20359/) (2025)
- [DiffuGR: Generative Document Retrieval with Diffusion Language Models](https://mlsanity.com/paper/2511.08150/) (2025)
- [HCMRM: A High-Consistency Multimodal Relevance Model for Search Ads](https://mlsanity.com/paper/2502.05822/) (2025)
- [Taming the One-Epoch Phenomenon in Online Recommendation System by Two-stage Contrastive ID Pre-training](https://mlsanity.com/paper/2508.18700/) (2025)
- [SUMMA: A Multimodal Large Language Model for Advertisement Summarization](https://mlsanity.com/paper/2508.20582/) (2025)
- [Applying Embedding-Based Retrieval to Airbnb Search](https://mlsanity.com/paper/2601.06873/) (2026)

## Abstract (arXiv)

Multimodal retrieval models are becoming increasingly important in scenarios such as food delivery, where rich multimodal features can meet diverse user needs and enable precise retrieval. Mainstream approaches typically employ a dual-tower architecture between queries and items, and perform joint optimization of intra-tower and inter-tower tasks. However, we observe that joint optimization often leads to certain modalities dominating the training process, while other modalities are neglected. In addition, inconsistent training speeds across modalities can easily result in the one-epoch problem. To address these challenges, we propose a staged pretraining strategy, which guides the model to focus on specialized tasks at each stage, enabling it to effectively attend to and utilize multimodal features, and allowing flexible control over the training process at each stage to avoid the one-epoch problem. Furthermore, to better utilize the semantic IDs that compress high-dimensional multimodal embeddings, we design both generative and discriminative tasks to help the model understand the associations between SIDs, queries, and item features, thereby improving overall performance. Extensive experiments on large-scale real-world Meituan data demonstrate that our method achieves improvements of 3.80%, 2.64%, and 2.17% on R@5, R@10, and R@20, and 5.10%, 4.22%, and 2.09% on N@5, N@10, and N@20 compared to mainstream baselines. Online A/B testing on the Meituan platform shows that our approach achieves a 1.12% increase in revenue and a 1.02% increase in click-through rate, validating the effectiveness and superiority of our method in practical applications.