DIRE: A Neural Approach to Decompiled Identifier Naming

Jeremy Lacomis, Pengcheng Yin, Edward J. Schwartz, Miltiadis Allamanis, Claire Le Goues, Graham Neubig, Bogdan VasilescuZenodo 2019 · DOI · OpenAlex

Abstract

This dataset is released as a companion to the paper "DIRE: A Neural Approach to Decompiled Identifier Naming", appearing in the proceedings of the 34th IEEE/ACM International Conference on Automated Software Engineering (ASE 2019). It contains information generated by decompiling 3,195,962 functions found in 164,632 unique binaries generated from C code scraped from GitHub. For practicality, the dataset is partitioned into 16 archives by the first hexadecimal digit of the SHA-256 hash of the binary used to generate it. Each of the 16 archives contains approximately 10,000 JSONL files, named according to a binary's hash. Each JSONL file consists of a single JSON object per-line corresponding to a single function in the decompiled binary. Archives are provided in both GZIP and BZIP2 format. See the README file for more information.
BibTeX
@inproceedings{lacomis2019dire,
  author = {Jeremy Lacomis and Pengcheng Yin and Edward J. Schwartz and Miltiadis Allamanis and Claire Le Goues and Graham Neubig and Bogdan Vasilescu},
  title = {{DIRE: A Neural Approach to Decompiled Identifier Naming}},
  booktitle = {Zenodo},
  year = {2019},
  doi = {10.5281/zenodo.3403078},
}

Loading related content...

Figure