# When Text-as-Vision Meets Semantic IDs in Generative Recommendation: An Empirical Study

arXiv: https://arxiv.org/abs/2601.14697 · HTML: https://mlsanity.com/paper/2601.14697/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 37 references that are in the map.
- First posted on arXiv 21 Jan 2026.

## Builds on (most cited first, 20 of 20)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [OneRec: Unifying Retrieve and Rank with Generative Recommender and Iterative Preference Alignment](https://mlsanity.com/paper/2502.18965/) (2025)
- [VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback](https://mlsanity.com/paper/1510.01784/) (2015)
- [Learning Vector-Quantized Item Representation for Transferable Sequential Recommenders](https://mlsanity.com/paper/2210.12316/) (2022)
- [Learnable Item Tokenization for Generative Recommendation](https://mlsanity.com/paper/2405.07314/) (2024)
- [Generative Recommendation: Towards Next-generation Recommender Paradigm](https://mlsanity.com/paper/2304.03516/) (2023)
- [Glyph: Scaling Context Windows via Visual-Text Compression](https://mlsanity.com/paper/2510.17800/) (2025)
- [Sparse Meets Dense: Unified Generative Recommendations with Cascaded Sparse-Dense Representations](https://mlsanity.com/paper/2503.02453/) (2025)
- [Generative Recommender with End-to-End Learnable Item Tokenization](https://mlsanity.com/paper/2409.05546/) (2024)
- [Order-agnostic Identifier for Large Language Model-based Generative Recommendation](https://mlsanity.com/paper/2502.10833/) (2025)
- [MMGRec: Multimodal Generative Recommendation with Transformer Model](https://mlsanity.com/paper/2404.16555/) (2024)
- [Enhancing Embedding Representation Stability in Recommendation Systems with Semantic ID](https://mlsanity.com/paper/2504.02137/) (2025)
- [FORGE: Forming Semantic Identifiers for Generative Retrieval in Industrial Datasets](https://mlsanity.com/paper/2509.20904/) (2025)
- [HiD-VAE: Interpretable Generative Recommendation via Hierarchical and Disentangled Semantic IDs](https://mlsanity.com/paper/2508.04618/) (2025)
- [MMQ: Multimodal Mixture-of-Quantization Tokenization for Semantic ID Generation and User Behavioral Adaptation](https://mlsanity.com/paper/2508.15281/) (2025)
- [Semantic Convergence: Harmonizing Recommender Systems via Two-Stage Alignment and Behavioral Semantic Tokenization](https://mlsanity.com/paper/2412.13771/) (2024)
- [See the Text: From Tokenization to Visual Reading](https://mlsanity.com/paper/2510.18840/) (2025)
- [DAS: Dual-Aligned Semantic IDs Empowered Industrial Recommender System](https://mlsanity.com/paper/2508.10584/) (2025)
- [Text or Pixels? It Takes Half: On the Token Efficiency of Visual Text Inputs in Multimodal LLMs](https://mlsanity.com/paper/2510.18279/) (2025)

## Built on by (most cited first, 3 of 3)

- [Hi-SAM: A Hierarchical Structure-Aware Multi-modal Framework for Large-Scale Recommendation](https://mlsanity.com/paper/2602.11799/) (2026)
- [Magnifying What Matters: Attention-Guided Adaptive Rendering for Visual Text Comprehension](https://mlsanity.com/paper/2606.12898/) (2026)
- [SIDScope: A Diagnostic Resource for Semantic-ID Interfaces in Generative Recommendation](https://mlsanity.com/paper/2608.18779/) (2026)

## Abstract (arXiv)

Semantic ID learning is a key interface in Generative Recommendation (GR) models, mapping items to discrete identifiers grounded in side information, most commonly via a pretrained text encoder. However, these text encoders are primarily optimized for well-formed natural language. In real-world recommendation data, item descriptions are often symbolic and attribute-centric, containing numerals, units, and abbreviations. These text encoders can break these signals into fragmented tokens, weakening semantic coherence and distorting relationships among attributes. Worse still, when moving to multimodal GR, relying on standard text encoders introduces an additional obstacle: text and image embeddings often exhibit mismatched geometric structures, making cross-modal fusion less effective and less stable. In this paper, we revisit representation design for Semantic ID learning by treating text as a visual signal. We conduct a systematic empirical study of OCR-based text representations, obtained by rendering item descriptions into images and encoding them with vision-based OCR models. Experiments across four datasets and two generative backbones show that OCR-text consistently matches or surpasses standard text embeddings for Semantic ID learning in both unimodal and multimodal settings. Furthermore, we find that OCR-based Semantic IDs remain robust under extreme spatial-resolution compression, indicating strong robustness and efficiency in practical deployments.