MoEKD: Mixture-of-Experts Knowledge Distillation for Robust and High-Performing Compressed Code Models
MoEKD:混合专家知识蒸馏用于鲁棒且高性能的压缩代码模型
专题命中 代码评测 :code model(title);分类 cs.SE
AI总结 本文提出MoEKD框架,通过混合专家架构实现多专家知识蒸馏,提升压缩模型的鲁棒性和性能,实验表明在漏洞检测任务中,MoEKD在对抗鲁棒性和预测性能上均优于现有基线方法。
Comments Accepted to the Research Track of the Evaluation and Assessment in Software Engineering (EASE) 2026