Large-Scale and Language-Oblivious Code Authorship Identification

Abuhamad, Mohammed; Abuhmed, Tamer; Mohaisen, Aziz; Nyang, DaeHun

doi:10.1145/3243734.3243738

Cited by 71 publications

(98 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We select models from recent work or use well-known models. For code comment generation and code authorship identification, we use recent work by Hu et al [20] and Abuhamad et al [21] respectively. For code clone detection, we use the well-documented baseline model, SourcererCC [22].…”

Section: Preliminariesmentioning

confidence: 99%

“…Techniques used in this task leverage features that express the programming style of programmers, such as layout and lexical features [42]. Previous works have found that the use of machine learning using TF-IDF features covering unigrams, bigrams and trigrams can be used to identify programmers at high accuracy [21].…”

Section: B Code Authorship Identificationmentioning

confidence: 99%

“…Dataset: As we did not manage to find any of the datasets used in prior work, we follow the same procedure described by Abuhamad et al [21] to build a similar dataset. This dataset is collected from program submissions to the Google Code Jam.…”

Section: B Code Authorship Identificationmentioning

confidence: 99%

“…On average, there are 106 lines in each Java program although the number of lines varies from 1 to over 70000 lines. Existing works on code authorship [21], [42] evaluate their approaches using accuracy. The dataset is constructed such that each author has the same number of programs in it.…”

Section: B Code Authorship Identificationmentioning

confidence: 99%

See 3 more Smart Citations

Assessing the Generalizability of Code2vec Token Embeddings

Kang

Bissyandé

2019

2019 34th IEEE/ACM International Conference on Automated Software Engineering (ASE)

View full text Add to dashboard Cite

Section: Preliminariesmentioning

confidence: 99%

Section: B Code Authorship Identificationmentioning

confidence: 99%

Section: B Code Authorship Identificationmentioning

confidence: 99%

Section: B Code Authorship Identificationmentioning

confidence: 99%

See 2 more Smart Citations

Assessing the Generalizability of Code2vec Token Embeddings

Kang

Bissyandé

2019

2019 34th IEEE/ACM International Conference on Automated Software Engineering (ASE)

View full text Add to dashboard Cite

“…In the dataset, a user's public data are the rating scores the user gave to mobile apps on Google Play, while the attribute is the city a user lives/lived political view, etc. Author identification attacks [11][12][13][14][15][16] Text document, program Author identity Website fingerprinting attacks [17][18][19][20][21] Network traffic Website…”

Section: Introductionmentioning

confidence: 99%

Defending Against Machine Learning Based Inference Attacks via Adversarial Examples: Opportunities and Challenges

Jia

Gong

2020

Adaptive Autonomous Secure Cyber Systems

View full text Add to dashboard Cite

As machine learning (ML) becomes more and more powerful and easily accessible, attackers increasingly leverage ML to perform automated large-scale inference attacks in various domains. In such an ML-equipped inference attack, an attacker has access to some data (called public data) of an individual, a software, or a system; and the attacker uses an ML classifier to automatically infer their private data. Inference attacks pose severe privacy and security threats to individuals and systems. Inference attacks are successful because private data are statistically correlated with public data, and ML classifiers can capture such statistical correlations. In this chapter, we discuss the opportunities and challenges of defending against ML-equipped inference attacks via adversarial examples. Our key observation is that attackers rely on ML classifiers in inference attacks. The adversarial machine learning community has demonstrated that ML classifiers have various vulnerabilities. Therefore, we can turn the vulnerabilities of ML into defenses against inference attacks. For example, ML classifiers are vulnerable to adversarial examples, which add carefully crafted noise to normal examples such that an ML classifier makes predictions for the examples as we desire. To defend against inference attacks, we can add carefully crafted noise into the public data to turn them into adversarial examples, such that attackers' classifiers make incorrect predictions for the private data. However, existing methods to construct adversarial examples are insufficient because they did not consider the unique challenges and requirements for the crafted noise at defending against inference attacks. In this chapter, we take defending against inference attacks in online social networks as an example to illustrate the opportunities and challenges.

show abstract