DualOptim: Enhancing Efficacy and Stability in Machine Unlearning with Dual Optimizers
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909879039426560 |
|---|---|
| author | Zhong, Xuyang Luo, Haochen Liu, Chen |
| author_facet | Zhong, Xuyang Luo, Haochen Liu, Chen |
| contents | Existing machine unlearning (MU) approaches exhibit significant sensitivity to hyperparameters, requiring meticulous tuning that limits practical deployment. In this work, we first empirically demonstrate the instability and suboptimal performance of existing popular MU methods when deployed in different scenarios. To address this issue, we propose Dual Optimizer (DualOptim), which incorporates adaptive learning rate and decoupled momentum factors. Empirical and theoretical evidence demonstrates that DualOptim contributes to effective and stable unlearning. Through extensive experiments, we show that DualOptim can significantly boost MU efficacy and stability across diverse tasks, including image classification, image generation, and large language models, making it a versatile approach to empower existing MU algorithms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_15827 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DualOptim: Enhancing Efficacy and Stability in Machine Unlearning with Dual Optimizers Zhong, Xuyang Luo, Haochen Liu, Chen Machine Learning Artificial Intelligence Existing machine unlearning (MU) approaches exhibit significant sensitivity to hyperparameters, requiring meticulous tuning that limits practical deployment. In this work, we first empirically demonstrate the instability and suboptimal performance of existing popular MU methods when deployed in different scenarios. To address this issue, we propose Dual Optimizer (DualOptim), which incorporates adaptive learning rate and decoupled momentum factors. Empirical and theoretical evidence demonstrates that DualOptim contributes to effective and stable unlearning. Through extensive experiments, we show that DualOptim can significantly boost MU efficacy and stability across diverse tasks, including image classification, image generation, and large language models, making it a versatile approach to empower existing MU algorithms. |
| title | DualOptim: Enhancing Efficacy and Stability in Machine Unlearning with Dual Optimizers |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2504.15827 |