Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.01343 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909644860948480 |
|---|---|
| author | Guo, Dongliang Hu, Mengxuan Guan, Zihan Hartvigsen, Thomas Li, Sheng |
| author_facet | Guo, Dongliang Hu, Mengxuan Guan, Zihan Hartvigsen, Thomas Li, Sheng |
| contents | Large multi-modal models inevitably decay over time as facts update and previously learned information becomes outdated. Traditional approaches such as fine-tuning are often impractical for updating these models due to their size and complexity. Instead, direct knowledge editing within the models presents a more viable solution. Current model editing techniques, however, typically overlook the unique influence ranges of different facts, leading to compromised model performance in terms of both generality and locality. To address this issue, we introduce the concept of the generality-locality trade-off in multi-modal model editing. We develop a new model editing dataset named OKEDIT, specifically designed to effectively evaluate this trade-off. Building on this foundation, we propose \textbf{BalancEdit}, a novel method for balanced model editing that dynamically achieves an optimal balance between generality and locality. BalancEdit utilizes a unique mechanism that generates both positive and negative samples for each fact to accurately determine its influence scope and incorporates these insights into the model's latent space using a discrete, localized codebook of edits, without modifying the underlying model weights. To our knowledge, this is the first approach explicitly addressing the generality-locality trade-off in multi-modal model editing. Our comprehensive results confirm the effectiveness of BalancEdit, demonstrating minimal trade-offs while maintaining robust editing capabilities. Our code and dataset are available at https://github.com/donglgcn/BalancEdit/tree/MMOKVQA. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_01343 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BalancEdit: Dynamically Balancing the Generality-Locality Trade-off in Multi-modal Model Editing Guo, Dongliang Hu, Mengxuan Guan, Zihan Hartvigsen, Thomas Li, Sheng Artificial Intelligence Large multi-modal models inevitably decay over time as facts update and previously learned information becomes outdated. Traditional approaches such as fine-tuning are often impractical for updating these models due to their size and complexity. Instead, direct knowledge editing within the models presents a more viable solution. Current model editing techniques, however, typically overlook the unique influence ranges of different facts, leading to compromised model performance in terms of both generality and locality. To address this issue, we introduce the concept of the generality-locality trade-off in multi-modal model editing. We develop a new model editing dataset named OKEDIT, specifically designed to effectively evaluate this trade-off. Building on this foundation, we propose \textbf{BalancEdit}, a novel method for balanced model editing that dynamically achieves an optimal balance between generality and locality. BalancEdit utilizes a unique mechanism that generates both positive and negative samples for each fact to accurately determine its influence scope and incorporates these insights into the model's latent space using a discrete, localized codebook of edits, without modifying the underlying model weights. To our knowledge, this is the first approach explicitly addressing the generality-locality trade-off in multi-modal model editing. Our comprehensive results confirm the effectiveness of BalancEdit, demonstrating minimal trade-offs while maintaining robust editing capabilities. Our code and dataset are available at https://github.com/donglgcn/BalancEdit/tree/MMOKVQA. |
| title | BalancEdit: Dynamically Balancing the Generality-Locality Trade-off in Multi-modal Model Editing |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2505.01343 |