Proj CDeepFuzz Paper Reading: COMET: Coverage-guided Model Generation For Deep Learning Library Testing

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
Github: https://github.com/maybeLee/COMET
Bug Type: Crash, NaN, inconsistency between the TensorFlow library and the ONNXRuntime library
Task: fuzzing API of DL Libraries
Method:

1. designs a set of mutation operators and a coverage-based search algorithm to diversify layer inputs, layer parameter values, and layer sequences in DL models
2. model synthesis

实验：
对象：ONNXRuntime, MXNet, Keras-MXNet, TF2ONNX, ONNX2PyTorch, Keras, TensorFlow, PyTorch
Competitors: Muffin, Lemon, Cradle
效果：

1. +32 bugs, 21 confirmed, 7 fixed

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
• 【五期李伟平】CCF-A（AAAI'21）Game of Gradients: Mitigating Irrelevant Clients in Federated Learning
• [论文阅读] Self-conditioned Image Generation via Generating Representations
• Learning Dynamic Query Combinations for Transformer-based Object** Detection and Segmentation论文阅读笔记
• 【略读论文|大模型相关】Zero-Shot Relational Learning on Temporal Knowledge Graphs with Large Language Models
• linux shell脚本for循环批量对bam文件构建索引并绘制geneBody coverage曲线
• sv coverage snippets
• 《An End-to-end Model for Entity-level Relation Extraction using Multi-instance Learning》阅读笔记
• CF940F Machine Learning题解
• CF940FMachine Learning题解

Coverage-guided Generation CDeepFuzz Coverage Learningcoverage-guided generation cdeepfuzz coverage coverage-guided diversifying generation cdeepfuzz inductive understanding pre-trained generation cdeepfuzz comprehensive cdeepfuzz compiler learning combinatorial cdeepfuzz learning reading deepmutation cdeepfuzz mutation learning large-scale tensorflow cdeepfuzz learning metamorphic cdeepfuzz compilers learning coverage