Dataflow Analysis-Inspired Deep Learning for Efficient Vulnerability Detection

Data package for "Dataflow Analysis-Inspired Deep Learning for Efficient Vulnerability Detection", published in ICSE 2024, with updates from Artifact Evaluation.
Paper link: https://www.computer.org/csdl/proceedings-article/icse/2024/021700a166/1RLIWqviwEM
See Github repo for updates: https://github.com/ISU-PAAL/DeepDFA
Data dictionary:
before.zip: CFGs of Big-Vul dataset, generated by Joern.preprocessed_data.zip: preprocessed data from Big-Vul for running DeepDFA, including preprocessed Joern CFGs and abstract dataflow embeddings.DeepDFA-code.zip: most recent version of the code as of the publication of this artifact, see Github repo for updates: https://github.com/ISU-PAAL/DeepDFAMSR_data_cleaned.csv: original Big-Vul dataset, see original source: https://github.com/ZeoVan/MSR_20_Code_vulnerability_CSV_DatasetMSR_LineVul: LineVul's preprocessed version of the Big-Vul dataset, see original source: https://github.com/awsm-research/LineVulChangelog:
v1 2023-09-20: original data package and Github repo published.v2 2024-01-04: added full instructions and bug fixes for Artifact Evaluation.v3 2024-01-10: integrated feedback from Artifact Evaluation.