ACE-Merging: Data-Free Model Merging with Adaptive Covariance Estimation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917391050473472 |
|---|---|
| author | Xu, Bo Wu, Haotian Lin, Hehai Huang, Weiquan Zhu, Beier Shu, Yao Qin, Chengwei |
| author_facet | Xu, Bo Wu, Haotian Lin, Hehai Huang, Weiquan Zhu, Beier Shu, Yao Qin, Chengwei |
| contents | Model merging aims to combine multiple task-specific expert models into a single model while preserving generalization across diverse tasks. However, interference among experts, especially when they are trained on different objectives, often leads to significant performance degradation. Despite recent progress, resolving this interference without data access, retraining, or architectural modification remains a fundamental challenge. This paper provides a theoretical analysis demonstrating that the input covariance of each task, which is a key factor for optimal merging, can be implicitly estimated from the parameter differences of its fine-tuned model, even in a fully data-free setting. Building on this insight, we introduce \acem, an Adaptive Covariance Estimation framework that effectively mitigates inter-task interference. Our approach features a principled, closed-form solution that contrasts with prior iterative or heuristic methods. Extensive experiments on both vision and language benchmarks demonstrate that \acem sets a new state-of-the-art among data-free methods. It consistently outperforms existing baselines; for example, \acem achieves an average absolute improvement of 4\% over the previous methods across seven tasks on GPT-2. Owing to its efficient closed-form formulation, \acem delivers superior performance with a modest computational cost, providing a practical and theoretically grounded solution for model merging. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_02945 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ACE-Merging: Data-Free Model Merging with Adaptive Covariance Estimation Xu, Bo Wu, Haotian Lin, Hehai Huang, Weiquan Zhu, Beier Shu, Yao Qin, Chengwei Computation and Language Model merging aims to combine multiple task-specific expert models into a single model while preserving generalization across diverse tasks. However, interference among experts, especially when they are trained on different objectives, often leads to significant performance degradation. Despite recent progress, resolving this interference without data access, retraining, or architectural modification remains a fundamental challenge. This paper provides a theoretical analysis demonstrating that the input covariance of each task, which is a key factor for optimal merging, can be implicitly estimated from the parameter differences of its fine-tuned model, even in a fully data-free setting. Building on this insight, we introduce \acem, an Adaptive Covariance Estimation framework that effectively mitigates inter-task interference. Our approach features a principled, closed-form solution that contrasts with prior iterative or heuristic methods. Extensive experiments on both vision and language benchmarks demonstrate that \acem sets a new state-of-the-art among data-free methods. It consistently outperforms existing baselines; for example, \acem achieves an average absolute improvement of 4\% over the previous methods across seven tasks on GPT-2. Owing to its efficient closed-form formulation, \acem delivers superior performance with a modest computational cost, providing a practical and theoretically grounded solution for model merging. |
| title | ACE-Merging: Data-Free Model Merging with Adaptive Covariance Estimation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2603.02945 |