Work

Paper vs Field Results: on the Impact of Eliminating Hindsight in Time-based (Vulnerability) Datasets for ML

Year: 2023

Type: conference-paper

Abstract: The Replication Package of "The Differentiated Replication Analysis on Software Vulnerability Detection with Time-based Datasets" Part 1 (NVD Vuldeepecker Dataset) The zipped package includes: Code that implements our methodology to generate the datasets Datasets that we created using our methodology and the original dataset (from NVD Vuldeepecker)- on_paper test = retrospective test- addNs test = believed_perspective test- field test = perspective test Pre-trained models that we generated during our evaluation (3 test results for each time point in the timeline [2008-2017]). Results of our evaluation. We also added the summarized result files (.xlsx) 1. Timeline of Datasets.xlsx 2. ML Evaluations.xlsx and the notebook to produce the charts in the paper: Charts.ipynb.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.8207883

Authors: Paramitha, Ranindya, Feng, Yuan, Massacci, Fabio

Institutions: University of Trento, Vrije Universiteit Amsterdam

Language: English

Topic: Software Engineering Research

Subfield: Information Systems

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funders: European Commission, Nederlandse Organisatie voor Wetenschappelijk Onderzoek

Awards: Theseus: Making patching happen, Assurance and certification in secure Multi-party Open Software and Services., Cybersecurity for AI-Augmented Systems, G7209445543