Revisiting RaBitQ and TurboQuant: A Symmetric Comparison of Methods, Theory, and Experiments
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914521591840768 |
|---|---|
| author | Gao, Jianyang Gou, Yutong Xu, Yuexuan Shi, Jifan Yang, Yongyi Li, Shuolin Wong, Raymond Chi-Wing Long, Cheng |
| author_facet | Gao, Jianyang Gou, Yutong Xu, Yuexuan Shi, Jifan Yang, Yongyi Li, Shuolin Wong, Raymond Chi-Wing Long, Cheng |
| contents | This technical note revisits the relationship between RaBitQ and TurboQuant under a unified comparison framework. We compare the two methods in terms of methodology, theoretical guarantees, and empirical performance, using a reproducible, transparent, and symmetric setup. Our results show that, despite the claimed advantage of TurboQuant, TurboQuant performs worse than RaBitQ in most tested settings of inner-product estimation, nearest-neighbor search and KV cache quantization. We further find that several reported runtime and recall results in the TurboQuant paper could not be reproduced from the released implementation under the stated configuration. Overall, this note clarifies the shared structure and genuine differences between the two lines of work, while documenting reproducibility issues in the experimental results reported by the TurboQuant paper. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_19528 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Revisiting RaBitQ and TurboQuant: A Symmetric Comparison of Methods, Theory, and Experiments Gao, Jianyang Gou, Yutong Xu, Yuexuan Shi, Jifan Yang, Yongyi Li, Shuolin Wong, Raymond Chi-Wing Long, Cheng Machine Learning Artificial Intelligence Databases This technical note revisits the relationship between RaBitQ and TurboQuant under a unified comparison framework. We compare the two methods in terms of methodology, theoretical guarantees, and empirical performance, using a reproducible, transparent, and symmetric setup. Our results show that, despite the claimed advantage of TurboQuant, TurboQuant performs worse than RaBitQ in most tested settings of inner-product estimation, nearest-neighbor search and KV cache quantization. We further find that several reported runtime and recall results in the TurboQuant paper could not be reproduced from the released implementation under the stated configuration. Overall, this note clarifies the shared structure and genuine differences between the two lines of work, while documenting reproducibility issues in the experimental results reported by the TurboQuant paper. |
| title | Revisiting RaBitQ and TurboQuant: A Symmetric Comparison of Methods, Theory, and Experiments |
| topic | Machine Learning Artificial Intelligence Databases |
| url | https://arxiv.org/abs/2604.19528 |