Unpaired Modality-Agnostic Generative Recommendation
2026-08-03 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors study a way to improve recommendation systems that generate suggestions using codes representing different types of data like images and text. They point out that current methods struggle when only one type of data (image or text) is available, making it hard to create consistent codes. To fix this, the authors propose UnpairGR, a new method that learns a shared code system from both paired and unpaired data, handling images and text more flexibly. Their approach uses shared model parts and separate input steps to keep things consistent and works well even when some data is missing. Experiments show that UnpairGR improves recommendation quality across different test datasets and scenarios.
Generative RecommendationAutoregressive GenerationSemantic IdentifiersMultimodal LearningTokenizationTransformerResidual CodebooksUnpaired DataRepresentation LearningCross-modal Consensus
Authors
Weihao Shen, Wei Chen, Fuwei Zhang, Meng Yuan, Yuqin Lan, Guojun Liu, Qingsong Hua, Wei Lin, Fuzhen Zhuang
Abstract
Generative Recommendation (GR) formulates recommendation as autoregressive generation over discrete semantic identifiers (IDs). Although recent multimodal GR methods improve semantic ID construction with visual and textual information, they typically require item-level paired observations, restricting tokenization to the intersection of modality availability. Moreover, incorporating unpaired observations is nontrivial because small representation shifts may cross quantization boundaries and produce incompatible identifier sequences. To address this challenge, we propose \textbf{Unpair}ed Modality-Agnostic \textbf{G}enerative \textbf{R}ecommendation (UnpairGR), which learns a unified semantic-ID space from paired, image-only, and text-only observations. UnpairGR confines modality-specific processing to lightweight input projections while sharing the subsequent Transformer and residual codebooks across all observation conditions. Paired observations establish a reliability-guided cross-modal consensus, whereas unimodal observations directly refine the same representations and codes. The learned tokenizer is then fixed to provide stationary targets for a single autoregressive recommender, without feature imputation, modality-specific codebooks, or fallback mappings. Extensive experiments on three benchmark datasets demonstrate that UnpairGR consistently improves recommendation performance under both fully observed and incomplete-observation settings.