English(EN) Quality-Driven Selective Mutation for Deep Learning

新框架通过选择高质量变异体来改进深度学习测试

作者 PulseAugur 编辑部 · [3 个来源] · 2026-04-24 15:15

研究人员开发了一个新的概率框架，用于评估深度学习测试中使用的变异体的质量。该框架根据抗拒性和真实性量化变异体质量，填补了当前深度学习文献中的空白。该方法允许对低质量变异算子配置进行排序和过滤，有可能将生成的变异体数量减少 55% 以上，同时保持其在测试和调试方面的有效性。 AI

影响引入了一种新方法来提高深度学习测试和调试的效率和有效性。

排序理由这是一篇介绍深度学习测试新框架的研究论文。

AI 生成摘要 · Google Gemini · 来自 3 个来源。我们如何撰写摘要 →

报道来源 [3]

arXiv cs.LG TIER_1 English(EN) · Zaheed Ahmed, Emmanuel Charleson Dapaah, Philip Makedonski, Jens Grabowski · 2026-04-27 04:00

面向深度学习的质量驱动选择性变异

arXiv:2604.22640v1 Announce Type: cross Abstract: Mutants support testing and debugging in two roles: (i) as test goals and (ii) as substitutes for real faults. Hard-to-kill mutants provide better guidance for test improvement, while realism is essential when mutants are used to …
arXiv cs.LG TIER_1 English(EN) · Jens Grabowski · 2026-04-24 15:15

面向深度学习的质量驱动选择性变异

Mutants support testing and debugging in two roles: (i) as test goals and (ii) as substitutes for real faults. Hard-to-kill mutants provide better guidance for test improvement, while realism is essential when mutants are used to simulate real bugs. Building on these roles, selec…
Hugging Face Daily Papers TIER_1 English(EN) · 2026-04-24 15:15

面向深度学习的质量驱动选择性变异

Mutants support testing and debugging in two roles: (i) as test goals and (ii) as substitutes for real faults. Hard-to-kill mutants provide better guidance for test improvement, while realism is essential when mutants are used to simulate real bugs. Building on these roles, selec…