FeatureSmith: Automatically Engineering Features for Malware Detection by Mining the Security Literature

By Ziyun Zhu, Tudor Dumitras. 2016

114 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Information Systems, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

1. Nicolas Papernot
2. Patrick Drew McDaniel
3. Jun Zhang
4. Ehab S. Al-Shaer
5. Michael Backes
6. Surya Nepal
7. Praveen Manoharan
8. Kathrin Grosse

In Information Systems

1. Patrick Drew McDaniel
2. Michael Backes
3. Surya Nepal
4. Praveen Manoharan
5. Nicolas Papernot
6. Jun Zhang
7. Kathrin Grosse
8. Yang Xiang

In Signal Processing

1. Nicolas Papernot
2. Patrick Drew McDaniel
3. Michael Backes
4. Surya Nepal
5. Kathrin Grosse
6. Praveen Manoharan
7. Yang Xiang
8. Jun Zhang

Citations in the corpus

• When Does Machine Learning FAIL? Generalized Transferability for Evasion and Poisoning Attacks (2018, neutral): “returned malicious labels. This is likely because antivirus products have started detecting them after the Drebin ground truth was labeled; incorrect labels in Drebin were also reported in prior work [56]. For the 11,847 benign-labeled base instances, 1,311 of their crafted instances were detected, ”
• Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning (2018, background): “Deep learning is ubiquitous: deep neural networks achieve exceptional performance on challenging tasks like machine translation [1], [2], diagnosing medical conditions such as diabetic retinopathy [3], [4] or pneumonia [5], malware detection [6], [7], [8], and classification of images [9], [10].”
• TTPDrill: Automatic and Accurate Extraction of Threat Actions from Unstructured Text of CTI Sources (2017, background): “[31], proposed the FeatureSmith, a system that extracts Android malware behaviors (SVOs) from scientic papers using typed dependency and part-of-speech.”
• Adversarial Examples for Malware Detection (2017, methodology): “[39], finally, applied feature selection and decision trees and achieved 1% false positives and 7.”
• A Survey of Android Malware Detection with Deep Neural Models (2020, background): “However, most of the ML-based methods require security experts to manually define the features to characterize Android malware, which heavily relies on experts’ experience, level of expertise, and the depth of domain knowledge [5, 6].”