Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2603.18492 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913025743650816 |
|---|---|
| author | Liu, Zongfang Tang, Shengkun Shen, Yifan Wang, Huan Yuan, Xin |
| author_facet | Liu, Zongfang Tang, Shengkun Shen, Yifan Wang, Huan Yuan, Xin |
| contents | Mixture-of-Experts (MoE) language models increase parameter capacity without proportional per-token compute, but the deployment still requires storing all experts, making expert pruning important for reducing memory and serving overhead. Existing task-agnostic expert pruning methods are typically calibration-dependent: they estimate expert importance from routing or activation statistics on a calibration set, which makes pruning outcomes sensitive to the choice of calibration set and adds substantial preprocessing cost. We introduce AIMER (\textbf{A}bsolute mean over root mean square \textbf{IM}portance for \textbf{E}xpert \textbf{R}anking), a simple calibration-free criterion that yields clear within-layer score separation and distinct expert stratification. Across 7B to 30B MoE language models at 25\% and 50\% pruning ratios over 16 benchmarks, AIMER consistently delivers competitive or stronger overall performance against state-of-the-art calibration-based expert pruning baselines with only 0.22--1.27 seconds for scoring the experts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_18492 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AIMER: Calibration-Free Task-Agnostic MoE Pruning Liu, Zongfang Tang, Shengkun Shen, Yifan Wang, Huan Yuan, Xin Machine Learning Mixture-of-Experts (MoE) language models increase parameter capacity without proportional per-token compute, but the deployment still requires storing all experts, making expert pruning important for reducing memory and serving overhead. Existing task-agnostic expert pruning methods are typically calibration-dependent: they estimate expert importance from routing or activation statistics on a calibration set, which makes pruning outcomes sensitive to the choice of calibration set and adds substantial preprocessing cost. We introduce AIMER (\textbf{A}bsolute mean over root mean square \textbf{IM}portance for \textbf{E}xpert \textbf{R}anking), a simple calibration-free criterion that yields clear within-layer score separation and distinct expert stratification. Across 7B to 30B MoE language models at 25\% and 50\% pruning ratios over 16 benchmarks, AIMER consistently delivers competitive or stronger overall performance against state-of-the-art calibration-based expert pruning baselines with only 0.22--1.27 seconds for scoring the experts. |
| title | AIMER: Calibration-Free Task-Agnostic MoE Pruning |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2603.18492 |