ESEC/FSE'23 Artifact for "NeuRI: Diversifying DNN Generation via Inductive Rule Inference"

Jiawei Liu, Jinjun Peng, Yuyao Wang, Lingming ZhangZenodo 2023 · DOI · OpenAlexGitHub resource

Abstract

This is the artifact for the ESEC/FSE'23 paper "NeuRI: Diversifying DNN Generation via Inductive Rule Inference". Deep Learning (DL) is prevalently used in various industries to improve decision-making and automate processes, driven by the ever-evolving DL libraries and compilers. The correctness of DL systems is crucial for trust in DL applications. <br> As such, the recent wave of research has been studying the automated synthesis of test-cases (i.e., DNN models and their inputs) for fuzzing DL systems. However, existing model generators only subsume a limited number of operators, lacking the ability to pervasively model operator constraints. <br> To address this challenge, we propose NeuRI, a fully automated approach for generating valid and diverse DL models composed of hundreds of types of operators. NeuRI adopts a three-step process: <br> (i) collecting valid and invalid API traces from various sources; <br> (ii) applying inductive program synthesis over the traces to infer the constraints for constructing valid models; and <br> (iii) using hybrid model generation which incorporates both symbolic and concrete operators. <br> Our evaluation shows that NeuRI improves branch coverage of TensorFlow and PyTorch by 24% and 15% over the state-of-the-art model-level fuzzers. NeuRI finds 100 new bugs for PyTorch and TensorFlow in four months, with 81 already fixed or confirmed. Of these, 9 bugs are labelled as high priority or security vulnerability, constituting 10% of all high-priority bugs of the period. <br> Open-source developers regard error-inducing tests reported by us as "high-quality" and "common in practice". For more information, please check the artifact GitHub repository: https://github.com/ise-uiuc/neuri-artifact
BibTeX
@inproceedings{liu2023esec,
  author = {Jiawei Liu and Jinjun Peng and Yuyao Wang and LINGMING ZHANG},
  title = {{ESEC/FSE'23 Artifact for "NeuRI: Diversifying DNN Generation via Inductive Rule Inference"}},
  booktitle = {Zenodo},
  year = {2023},
  doi = {10.5281/zenodo.8319847},
}

Loading related content...

Figure