EMBER: An Open Dataset for Training Static PE Malware Machine Learning Models

Abstract

This paper describes EMBER: a labeled benchmark dataset for training machine learning models to statically detect malicious Windows portable executable files. The dataset includes features extracted from 1.1M binary files: 900K training samples (300K malicious, 300K benign, 300K unlabeled) and 200K test samples (100K malicious, 100K benign). To accompany the dataset, we also release open source code for extracting features from additional binaries so that additional sample features can be appended to the dataset. This dataset fills a void in the information security machine learning community: a benign/malicious dataset that is large, open and general enough to cover several interesting use cases. We enumerate several use cases that we considered when structuring the dataset. Additionally, we demonstrate one use case wherein we compare a baseline gradient boosted decision tree model trained using LightGBM with default settings to MalConv, a recently published end-to-end (featureless) deep learning model for malware detection. Results show that even without hyper-parameter optimization, the baseline EMBER model outperforms MalConv. The authors hope that the dataset, code and baseline model provided by EMBER will help invigorate machine learning research for malware detection, in much the same way that benchmark datasets have advanced computer vision research.

References in corpus (2)

• Microsoft Malware Classification Challenge
• Deep Neural Network Based Malware Detection Using Two Dimensional Binary Program Features

Cited by in corpus (32)

• Adversarial Attacks against Windows PE Malware Detection: A Survey of the State-of-the-Art
• Malware Makeover: Breaking ML-based Static Analysis by Modifying Executable Bytes
• Explanation-Guided Backdoor Poisoning Attacks Against Malware Classifiers
• A Survey on Cross-Architectural IoT Malware Threat Hunting
• Machine Learning in Cyber-Security - Problems, Challenges and Data Sets
• Machine Learning (In) Security: A Stream of Problems
• A Survey of Machine Learning Methods and Challenges for Windows Malware Classification
• MalPhase: Fine-Grained Malware Detection Using Network Flow Data
• Explaining Vulnerabilities of Deep Learning to Adversarial Malware Binaries
• Virus-MNIST: A Benchmark Malware Dataset
• MAB-Malware: A Reinforcement Learning Framework for Attacking Static Malware Classifiers
• Performance analysis of Machine learning algorithms for predicting malware
• A Framework for Cluster and Classifier Evaluation in the Absence of Reference Labels
• Leveraging Uncertainty for Improved Static Malware Detection Under Extreme False Positive Constraints
• Stack-based Buffer Overflow Detection using Recurrent Neural Networks
• Learning from Context: Exploiting and Interpreting File Path Information for Better Malware Detection
• Malware Detection Using Frequency Domain-Based Image Visualization and Deep Learning
• Advances In Malware Detection- An Overview
• Towards an Automated Pipeline for Detecting and Classifying Malware through Machine Learning
• Adversarial Transfer Attacks With Unknown Data and Class Overlap
• Adversarial Robustness with Non-uniform Perturbations
• Partitioned Learned Bloom Filter
• Evading Malware Classifiers via Monte Carlo Mutant Feature Discovery
• Neurlux: Dynamic Malware Analysis Without Feature Engineering
• Training Transformers for Information Security Tasks: A Case Study on Malicious URL Prediction
• Gradient-based Data Subversion Attack Against Binary Classifiers
• Being Single Has Benefits. Instance Poisoning to Deceive Malware Classifiers
• Mind the Gap: On Bridging the Semantic Gap between Machine Learning and Information Security
• Mill.jl and JsonGrinder.jl: automated differentiable feature extraction for learning from raw JSON data
• Dissecting Malware in the Wild
• A Comparison of State-of-the-Art Techniques for Generating Adversarial Malware Binaries
• An MDL-Based Classifier for Transactional Datasets with Application in Malware Detection