# MMGRec: Multimodal Generative Recommendation with Transformer Model

arXiv: https://arxiv.org/abs/2404.16555 · HTML: https://mlsanity.com/paper/2404.16555/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Apr 2024.

## Built on by (most cited first, 20 of 20)

- [Sparse Meets Dense: Unified Generative Recommendations with Cascaded Sparse-Dense Representations](https://mlsanity.com/paper/2503.02453/) (2025)
- [Unifying Generative and Dense Retrieval for Sequential Recommendation](https://mlsanity.com/paper/2411.18814/) (2024)
- [ActionPiece: Contextually Tokenizing Action Sequences for Generative Recommendation](https://mlsanity.com/paper/2502.13581/) (2025)
- [Multimodal Quantitative Language for Generative Recommendation](https://mlsanity.com/paper/2504.05314/) (2025)
- [Node Identifiers: Compact, Discrete Representations for Efficient Graph Learning](https://mlsanity.com/paper/2405.16435/) (2024)
- [Unleash LLMs Potential for Sequential Recommendation by Coordinating Dual Dynamic Index Mechanism](https://mlsanity.com/paper/2409.09253/) (2024)
- [IGD: Token Decisiveness Modeling via Information Gain in LLMs for Personalized Recommendation](https://mlsanity.com/paper/2506.13229/) (2025)
- [A Simple Contrastive Framework Of Item Tokenization For Generative Recommendation](https://mlsanity.com/paper/2506.16683/) (2025)
- [GRACE: Generative Recommendation via Journey-Aware Sparse Attention on Chain-of-Thought Tokenization](https://mlsanity.com/paper/2507.14758/) (2025)
- [Pctx: Tokenizing Personalized Context for Generative Recommendation](https://mlsanity.com/paper/2510.21276/) (2025)
- [Multi-Aspect Cross-modal Quantization for Generative Recommendation](https://mlsanity.com/paper/2511.15122/) (2025)
- [When Text-as-Vision Meets Semantic IDs in Generative Recommendation: An Empirical Study](https://mlsanity.com/paper/2601.14697/) (2026)
- [Dreaming User Multimodal Representation Guided by The Platonic Representation Hypothesis for Micro-Video Recommendation](https://mlsanity.com/paper/2410.03538/) (2024)
- [FOUNDv2: Learning Unified User Quantized Tokenizers for User Representation](https://mlsanity.com/paper/2508.00956/) (2025)
- [Multimodal Representation Learning Conditioned on Semantic Relations](https://mlsanity.com/paper/2508.17497/) (2025)
- [Learning Variable-Length Tokenization for Generative Recommendation](https://mlsanity.com/paper/2605.17779/) (2026)
- [SynGR: Unleashing the Potential of Cross-Modal Synergy for Generative Recommendation](https://mlsanity.com/paper/2605.18920/) (2026)
- [GTS-LUM: Reshaping User Behavior Modeling with LLMs in Telecommunications Industry](https://mlsanity.com/paper/2504.06511/) (2025)
- [Q-BERT4Rec: Quantized Semantic-ID Representation Learning for Multimodal Recommendation](https://mlsanity.com/paper/2512.02474/) (2025)
- [HASTE: Hierarchical and Attribute-aware Supervision for Generative Recommendation](https://mlsanity.com/paper/2610.17368/) (2026)

## Abstract (arXiv)

Multimodal recommendation aims to recommend user-preferred candidates based on her/his historically interacted items and associated multimodal information. Previous studies commonly employ an embed-and-retrieve paradigm: learning user and item representations in the same embedding space, then retrieving similar candidate items for a user via embedding inner product. However, this paradigm suffers from inference cost, interaction modeling, and false-negative issues. Toward this end, we propose a new MMGRec model to introduce a generative paradigm into multimodal recommendation. Specifically, we first devise a hierarchical quantization method Graph RQ-VAE to assign Rec-ID for each item from its multimodal and CF information. Consisting of a tuple of semantically meaningful tokens, Rec-ID serves as the unique identifier of each item. Afterward, we train a Transformer-based recommender to generate the Rec-IDs of user-preferred items based on historical interaction sequences. The generative paradigm is qualified since this model systematically predicts the tuple of tokens identifying the recommended item in an autoregressive manner. Moreover, a relation-aware self-attention mechanism is devised for the Transformer to handle non-sequential interaction sequences, which explores the element pairwise relation to replace absolute positional encoding. Extensive experiments evaluate MMGRec's effectiveness compared with state-of-the-art methods.