Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2404.13322 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912169629581312 |
|---|---|
| author | Li, Kunxi Zhan, Tianyu Fu, Kairui Zhang, Shengyu Kuang, Kun Li, Jiwei Zhao, Zhou Wu, Fan Wu, Fei |
| author_facet | Li, Kunxi Zhan, Tianyu Fu, Kairui Zhang, Shengyu Kuang, Kun Li, Jiwei Zhao, Zhou Wu, Fan Wu, Fei |
| contents | In this study, we focus on heterogeneous knowledge transfer across entirely different model architectures, tasks, and modalities. Existing knowledge transfer methods (e.g., backbone sharing, knowledge distillation) often hinge on shared elements within model structures or task-specific features/labels, limiting transfers to complex model types or tasks. To overcome these challenges, we present MergeNet, which learns to bridge the gap of parameter spaces of heterogeneous models, facilitating the direct interaction, extraction, and application of knowledge within these parameter spaces. The core mechanism of MergeNet lies in the parameter adapter, which operates by querying the source model's low-rank parameters and adeptly learning to identify and map parameters into the target model. MergeNet is learned alongside both models, allowing our framework to dynamically transfer and adapt knowledge relevant to the current stage, including the training trajectory knowledge of the source model. Extensive experiments on heterogeneous knowledge transfer demonstrate significant improvements in challenging settings, where representative approaches may falter or prove less applicable. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_13322 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MergeNet: Knowledge Migration across Heterogeneous Models, Tasks, and Modalities Li, Kunxi Zhan, Tianyu Fu, Kairui Zhang, Shengyu Kuang, Kun Li, Jiwei Zhao, Zhou Wu, Fan Wu, Fei Machine Learning Artificial Intelligence In this study, we focus on heterogeneous knowledge transfer across entirely different model architectures, tasks, and modalities. Existing knowledge transfer methods (e.g., backbone sharing, knowledge distillation) often hinge on shared elements within model structures or task-specific features/labels, limiting transfers to complex model types or tasks. To overcome these challenges, we present MergeNet, which learns to bridge the gap of parameter spaces of heterogeneous models, facilitating the direct interaction, extraction, and application of knowledge within these parameter spaces. The core mechanism of MergeNet lies in the parameter adapter, which operates by querying the source model's low-rank parameters and adeptly learning to identify and map parameters into the target model. MergeNet is learned alongside both models, allowing our framework to dynamically transfer and adapt knowledge relevant to the current stage, including the training trajectory knowledge of the source model. Extensive experiments on heterogeneous knowledge transfer demonstrate significant improvements in challenging settings, where representative approaches may falter or prove less applicable. |
| title | MergeNet: Knowledge Migration across Heterogeneous Models, Tasks, and Modalities |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2404.13322 |