一句话：提出互补集引导的约束精化方法，用遗传算法提升深度学习库测试约束完整性。
深度学习库在 AI 系统中非常重要，近期许多工作通过将张量操作的输入建模为约束来指导测试用例生成，以保证其可靠性。然而，这些约束可能缩小搜索空间，导致测试不完整。本文提出互补集引导的精化方法，其基本思想是检查约束的互补集是否能产生有效测试用例；如果能，则原约束不完整，需要精化。基于该思想，作者设计了自动约束精化工具 DeepConstr，采用遗传算法提升约束的完整性。在 PyTorch 和 TensorFlow 上评估，DeepConstr 发现 84 个未知 bug，其中 72 个被确认、51 个被修复。与最新 fuzzer 相比，DeepConstr 在 NNSmith 支持的算子中提升了 43.44% 的覆盖率，在 NeuRI 支持的算子中提升了 59.16%。
Deep learning library is important in AI systems. Recently, many works have been proposed to ensure its reliability. They often model inputs of tensor operations as constraints to guide the generation of test cases. However, these constraints may narrow the search space, resulting in incomplete testing. This paper introduces a complementary set-guided refinement that can enhance the completeness of constraints. The basic idea is to see if the complementary set of constraints yields valid test cases. If so, the original constraint is incomplete and needs refinement. Based on this idea, we design an automatic constraint refinement tool, DeepConstr, which adopts a genetic algorithm to refine constraints for better completeness. We evaluated it on two DL libraries, PyTorch and TensorFlow. DeepConstr discovered 84 unknown bugs, out of which 72 were confirmed, with 51 fixed. Compared to state-of-the-art fuzzers, DeepConstr increased coverage for 43.44% of operators supported by NNSmith, and 59.16% of operators supported by NeuRI.
DOI 原文 ·
@inproceedings{paperbot8613,
title = {Towards More Complete Constraints for Deep Learning Library Testing via Complementary Set Guided Refinement},
author = {Gwihwan Go and Chijin Zhou and Quan Zhang and Xiazijian Zou and Heyuan Shi and Yu Jiang},
booktitle = {ISSTA},
year = {2024},
doi = {10.1145/3650212.3680364}
}