上一页 1 2 3 4 5 6 7 ··· 79 下一页
2023年9月7日
Proj CDeepFuzz Paper Reading: Software Testing with Large Language Model: Survey, Landscape, and Vision
摘要： ## Abstract 本文: Task: Review on the use of LLMs in software testing Method: 1. analyzes 52 relevant studies ## 1. Intro ![](https://img2023.cnblogs.co 阅读全文
posted @ 2023-09-07 22:43 雪溯 阅读(53) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: PELICAN: Exploiting Backdoors of Naturally Trained Deep Learning Models In Binary Code Analysis
摘要： ## Abstract 背景： 1. 本文研究的不是被恶意植入的后门，而是products of defects in training 2. 攻击模式: injecting some small fixed input pattern(backdoor) to induce misclassifi 阅读全文
posted @ 2023-09-07 18:13 雪溯 阅读(19) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: Decompiling x86 Deep Neural Network Executables
摘要： ## Abstract 本文: BTD github: https://github.com/monkbai/DNN-decompiler/ Task: a decompiler for DNN models to output DNN specifications including: opera 阅读全文
posted @ 2023-09-07 17:32 雪溯 阅读(12) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: Balancing Effectiveness and Flakiness of Non-Deterministic Machine Learning Tests
摘要： ## Abstract 背景：In fact, some of the latest findings suggest that the existence of adversarial attacks may be an inherent weakness of deep learning mod 阅读全文
posted @ 2023-09-07 08:01 雪溯 阅读(14) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: NeuRI: Diversifying DNN Generation via Inductive Rule Inference
摘要： ## Abstract 背景：The correctness of DL systems is crucial for trust in DL applications 本文: NeuRI BaseTool: FreeFuzz Github: https://github.com/ise-uiuc/ 阅读全文
posted @ 2023-09-07 06:38 雪溯 阅读(23) 评论(0) 推荐(0) 编辑
2023年9月6日
Proj CDeepFuzz Paper Reading: Natural attack for pre-trained models of code
摘要： ## Abstract 背景：目前大多数的adversarial attack method on pre-trained models of code忽略了perturbations should be natural to human judges(naturalness requirement 阅读全文
posted @ 2023-09-06 23:34 雪溯 阅读(20) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: COMET: Coverage-guided Model Generation For Deep Learning Library Testing
摘要： ## Abstract 背景：已有的方法(Muffin, Lemon, Cradle) can cover at most 34.1% layer inputs, 25.9% layer parameter values, and 15.6% layer sequences. 本文：COMET Gi 阅读全文
posted @ 2023-09-06 23:03 雪溯 阅读(22) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: IvySyn: Automated Vulnerability Discovery in Deep Learning Frameworks
摘要： ## Abstract 本文：IvySyn Task: discover memory error vulnerabilities in DL frameworks BugType: memory safety errors, fatal runtime errors Method: 1. 利用na 阅读全文
posted @ 2023-09-06 21:34 雪溯 阅读(19) 评论(0) 推荐(0) 编辑
2023年9月5日
Proj CDeepFuzz Paper Reading: Invariance-inducing regularization using worst-case transformations suffices to boost accuracy and spatial robustness
摘要： ## Abstract 本文： Task: 1. prove invariance-inducing regularizers can increase predictive accuracy for worst-case spatial transformations 2. prove that 阅读全文
posted @ 2023-09-05 23:14 雪溯 阅读(4) 评论(0) 推荐(0) 编辑
Proj CDeepFuzz Paper Reading: Framework for Evaluating Faithfulness of Local Explanations
摘要： ## Abstract 本文: Task: 1. study the faithfulness of an explanation system to the underlying prediction model on consistency and sufficiency 2. introduc 阅读全文
posted @ 2023-09-05 22:00 雪溯 阅读(5) 评论(0) 推荐(0) 编辑
上一页 1 2 3 4 5 6 7 ··· 79 下一页