Lifelong Knowledge Editing for Vision Language Models with Low-Rank Mixture-of-Experts
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912274056216576 |
|---|---|
| author | Chen, Qizhou Wang, Chengyu Wang, Dakan Zhang, Taolin Li, Wangyue He, Xiaofeng |
| author_facet | Chen, Qizhou Wang, Chengyu Wang, Dakan Zhang, Taolin Li, Wangyue He, Xiaofeng |
| contents | Model editing aims to correct inaccurate knowledge, update outdated information, and incorporate new data into Large Language Models (LLMs) without the need for retraining. This task poses challenges in lifelong scenarios where edits must be continuously applied for real-world applications. While some editors demonstrate strong robustness for lifelong editing in pure LLMs, Vision LLMs (VLLMs), which incorporate an additional vision modality, are not directly adaptable to existing LLM editors. In this paper, we propose LiveEdit, a LIfelong Vision language modEl Edit to bridge the gap between lifelong LLM editing and VLLMs. We begin by training an editing expert generator to independently produce low-rank experts for each editing instance, with the goal of correcting the relevant responses of the VLLM. A hard filtering mechanism is developed to utilize visual semantic knowledge, thereby coarsely eliminating visually irrelevant experts for input queries during the inference stage of the post-edited model. Finally, to integrate visually relevant experts, we introduce a soft routing mechanism based on textual semantic relevance to achieve multi-expert fusion. For evaluation, we establish a benchmark for lifelong VLLM editing. Extensive experiments demonstrate that LiveEdit offers significant advantages in lifelong VLLM editing scenarios. Further experiments validate the rationality and effectiveness of each module design in LiveEdit. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_15432 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Lifelong Knowledge Editing for Vision Language Models with Low-Rank Mixture-of-Experts Chen, Qizhou Wang, Chengyu Wang, Dakan Zhang, Taolin Li, Wangyue He, Xiaofeng Computation and Language Computer Vision and Pattern Recognition Model editing aims to correct inaccurate knowledge, update outdated information, and incorporate new data into Large Language Models (LLMs) without the need for retraining. This task poses challenges in lifelong scenarios where edits must be continuously applied for real-world applications. While some editors demonstrate strong robustness for lifelong editing in pure LLMs, Vision LLMs (VLLMs), which incorporate an additional vision modality, are not directly adaptable to existing LLM editors. In this paper, we propose LiveEdit, a LIfelong Vision language modEl Edit to bridge the gap between lifelong LLM editing and VLLMs. We begin by training an editing expert generator to independently produce low-rank experts for each editing instance, with the goal of correcting the relevant responses of the VLLM. A hard filtering mechanism is developed to utilize visual semantic knowledge, thereby coarsely eliminating visually irrelevant experts for input queries during the inference stage of the post-edited model. Finally, to integrate visually relevant experts, we introduce a soft routing mechanism based on textual semantic relevance to achieve multi-expert fusion. For evaluation, we establish a benchmark for lifelong VLLM editing. Extensive experiments demonstrate that LiveEdit offers significant advantages in lifelong VLLM editing scenarios. Further experiments validate the rationality and effectiveness of each module design in LiveEdit. |
| title | Lifelong Knowledge Editing for Vision Language Models with Low-Rank Mixture-of-Experts |
| topic | Computation and Language Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2411.15432 |