Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2503.19868 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915328683933696 |
|---|---|
| author | Kim, Sungyeon Zhu, Xinliang Lin, Xiaofan Bastan, Muhammet Gray, Douglas Kwak, Suha |
| author_facet | Kim, Sungyeon Zhu, Xinliang Lin, Xiaofan Bastan, Muhammet Gray, Douglas Kwak, Suha |
| contents | Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_19868 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GENIUS: A Generative Framework for Universal Multimodal Search Kim, Sungyeon Zhu, Xinliang Lin, Xiaofan Bastan, Muhammet Gray, Douglas Kwak, Suha Information Retrieval Artificial Intelligence Computer Vision and Pattern Recognition Machine Learning Generative retrieval is an emerging approach in information retrieval that generates identifiers (IDs) of target data based on a query, providing an efficient alternative to traditional embedding-based retrieval methods. However, existing models are task-specific and fall short of embedding-based retrieval in performance. This paper proposes GENIUS, a universal generative retrieval framework supporting diverse tasks across multiple modalities and domains. At its core, GENIUS introduces modality-decoupled semantic quantization, transforming multimodal data into discrete IDs encoding both modality and semantics. Moreover, to enhance generalization, we propose a query augmentation that interpolates between a query and its target, allowing GENIUS to adapt to varied query forms. Evaluated on the M-BEIR benchmark, it surpasses prior generative methods by a clear margin. Unlike embedding-based retrieval, GENIUS consistently maintains high retrieval speed across database size, with competitive performance across multiple benchmarks. With additional re-ranking, GENIUS often achieves results close to those of embedding-based methods while preserving efficiency. |
| title | GENIUS: A Generative Framework for Universal Multimodal Search |
| topic | Information Retrieval Artificial Intelligence Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2503.19868 |