Replication package for the paper Featherweight Assisted Vulnerability Discovery

This repository contains the replication package for the paper "Featherweight Assisted Vulnerability Discovery", David Binkley, Leon Moonen, Sibren Isaacman, Information and Software Technology, 2022, 106844, ISSN 0950-5849, DOI: 10.1016/j.infsof.2022.106844. https://www.sciencedirect.com/science/article/pii/S0950584922000209. The source code is distributed under the MIT license, the data is distributed under the CC BY 4.0 license. Repository Organization The overall process consists of three steps, organized as three directories: gathering of the labeled function names that are used as the source for step 2, in names dangerous word identification, in dangerous-words analysis of the data gathered during step 2, in analysis The directory Model holds a copy of the pre-trained LAVDNN model as provided by the authors at https://github.com/StablelJay/LAVDNN/raw/master/Model/model_of_LAVDNN Requirements The following tools are required for the replication: python >= 3.5 R tcsh csvcut from csvkit cntk as keras backend for running the LAVDNN model In addition, the following python packages are needed pandas, h5py, requests, and lizard for the collection of our input data numpy and keras for dangerous word identification Finally, for the analysis in step 3, the following R libraries are needed: agricolae, ggplot2, reshape2, xtable Citation If you build on this data or code, please cite this work by referring to the paper: @article{binkley2022:featherweight, title = {Featherweight assisted vulnerability discovery}, author = {David Binkley and Leon Moonen and Sibren Isaacman}, journal = {Information and Software Technology}, pages = {106844}, year = {2022}, issn = {0950-5849}, doi = {https://doi.org/10.1016/j.infsof.2022.106844}, url = {https://www.sciencedirect.com/science/article/pii/S0950584922000209}, copyright = {Open Access}, publisher = {Elsevier}, } The GitHub repository containing the code to automatically collect the dataset and replicate the experiments can be found at https://github.com/secureIT-project/FAVDPart of this work has been financially supported by the Research Council of Norway through the secureIT project (RCN contract #288787)