Proj CDeepFuzz Paper Reading: ACETest: Automated Constraint Extraction for Testing Deep Learning Operators

Abstract

Github: https://github.com/shijy16/ACETest
背景：

1. DL operators 用来计算多维tensors，很重要

本文：ACETest
Task: automatically extract input validation constraints from the code to test DL operators
方法：identify the input validation code，据此生成test cases
Bug Types: Crash, Timeout?
实验：
对象：Tensorflow, PyTorch
Competitors: FreeFuzz, DeepREL, DocTer
效果：

1. 比目前方法质量高
2. +96.4%constraints
3. +1.95-55 times bugs
4. +18 unknown bugs, 87 confirmed, 5 CVEs

本栏目推荐文章
• 《RAPL: A Relation-Aware Prototype Learning Approach for Few-Shot Document-Level Relation Extraction》阅读笔记
• 《Span-Based Joint Entity and Relation Extraction with Transformer Pre-Training》阅读笔记
• 《An End-to-end Model for Entity-level Relation Extraction using Multi-instance Learning》阅读笔记
• 差分约束(Differential constraint)
• 《A Novel Table-to-Graph Generation Approach for Document-Level Joint Entity and Relation Extraction》阅读笔记
• Go - Keywords, Operators and Punctuation
• codeblock快捷键+VS code快捷键+DW_minmax文件+莱文斯坦距离+char* 和 char[]区别+可增删constraint的实现方法+$fatal的传参+巴科斯范式命令行语法+ralgen生成的寄存器模型存在覆盖率的警告（uvm-1.2）
• Graph regularized non-negative matrix factorization with prior knowledge consistency constraint for drug-target interactions prediction
• 《REBEL Relation Extraction By End-to-end Language generation》阅读笔记
• python-task3：Data Types and Operators

Constraint Extraction CDeepFuzz Automated Operatorsconstraint extraction cdeepfuzz automated vulnerability frameworks cdeepfuzz automated deepxplore cdeepfuzz automated whitebox operators automated punctuation operators keywords and operators utility rxjs mathematical aggregate operators rxjs combinelatest node_modules operators expected extraction