mlsanityData to 14 Oct 2026

Assemblage: Automatic Binary Dataset Construction for Machine Learning

arXiv 2405.03991 ↗First posted 7 May 2024cs.CR cs.LGcited by 12

Not on the map. Why?

Abstract (arXiv)

Binary code is pervasive, and binary analysis is a key task in reverse engineering, malware classification, and vulnerability discovery. Unfortunately, while there exist large corpora of malicious binaries, obtaining high-quality corpora of benign binaries for modern systems has proven challenging (e.g., due to licensing issues). Consequently, machine learning based pipelines for binary analysis utilize either costly commercial corpora (e.g., VirusTotal) or open-source binaries (e.g., coreutils) available in limited quantities. To address these issues, we present Assemblage: an extensible cloud-based distributed system that crawls, configures, and builds Windows PE binaries to obtain high-quality binary corpuses suitable for training state-of-the-art models in binary analysis. We have run Assemblage on AWS over the past year, producing 890k Windows PE and 428k Linux ELF binaries across 29 configurations. Assemblage is designed to be both reproducible and extensible, enabling users to publish "recipes" for their datasets, and facilitating the extraction of a wide array of features. We evaluated Assemblage by using its data to train modern learning-based pipelines for compiler provenance and binary function similarity. Our results illustrate the practical need for robust corpora of high-quality Windows PE binaries in training modern learning-based binary analyses. Assemblage code is open sourced under the MIT license, and the dataset can be downloaded from https://assemblage-dataset.net

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Builds on 36 of its 71 references that are in the map. First posted on arXiv 7 May 2024.

Builds on

36 of 71 references.

References · 36

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2023GPT-4 Technical Reportcited by 18,396
  4. 2022Emergent Abilities of Large Language Modelscited by 2,512
  5. 2019Graph Matching Networks for Learning the Similarity of Graph Structured Objectscited by 205
  6. 2018EMBER: An Open Dataset for Training Static PE Malware Machine Learning Modelscited by 165
  7. 2019A Step Toward Quantifying Independently Reproducible Machine Learning Researchcited by 77
  8. 2020SOREL-20M: A Large Scale Benchmark Dataset for Malicious PE Detectioncited by 49
  9. 2021PalmTree: Learning an Assembly Language Model for Instruction Embeddingcited by 49
  10. 2022jTrans: Jump-Aware Transformer for Binary Code Similaritycited by 44
  11. 2024LLM4Decompile: Decompiling Binary Code with Large Language Modelscited by 39
  12. 2019DIRE: A Neural Approach to Decompiled Identifier Namingcited by 36
24 more
  1. 2020A Survey of Machine Learning Methods and Challenges for Windows Malware Classificationcited by 33
  2. 2023Extending Source Code Pre-Trained Language Models to Summarise Decompiled Binariescited by 26
  3. 2021Augmenting Decompiler Output with Learned Variable Names and Typescited by 23
  4. 2020XDA: Accurate, Robust Disassembly with Transfer Learningcited by 21
  5. 2020Automatic Yara Rule Generation Using Biclusteringcited by 17
  6. 2021MOTIF: A Large Malware Reference Dataset with Ground Truth Family Labelscited by 17
  7. 2020Research Reproducibility as a Survival Analysiscited by 16
  8. 2021A Framework for Cluster and Classifier Evaluation in the Absence of Reference Labelscited by 16
  9. 2020SourceFinder: Finding Malware Source-Code from Publicly Available Repositoriescited by 15
  10. 2020Revisiting Binary Code Similarity Analysis using Interpretable Feature Engineering and Lessons Learnedcited by 14
  11. 2023kTrans: Knowledge-Aware Transformer for Binary Code Embeddingcited by 10
  12. 2018Static Malware Detection & Subterfuge: Quantifying the Robustness of Machine Learning and Current Anti-Viruscited by 8
  13. 2019I-MAD: Interpretable Malware Detector Using Galaxy Transformercited by 8
  14. 2022BinBert: Binary Code Understanding with a Fine-tunable and Execution-aware Transformercited by 8
  15. 2021Rank-1 Similarity Matrix Decomposition For Modeling Changes in Antivirus Consensus Through Timecited by 7
  16. 2021Semantic-aware Binary Code Representation with BERTcited by 6
  17. 2020paper2repo: GitHub Repository Recommendation for Academic Paperscited by 5
  18. 2023AVScan2Vec: Feature Learning on Antivirus Scan Data for Production-Scale Malware Corporacited by 5
  19. 2023Small Effect Sizes in Malware Detection? Make Harder Train/Test Splits!cited by 4
  20. 2022UniASM: Binary Code Similarity Detection without Fine-tuningcited by 3
  21. 2018Ubuntu One Investigation: Detecting Evidences on Client Machinescited by 2
  22. 2021Evading Malware Classifiers via Monte Carlo Mutant Feature Discoverycited by 2
  23. 2022Accelerating Frank-Wolfe via Averaging Step Directionscited by 2
  24. 2024Synthetic Datasets for Program Similarity Researchcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 2; 2025: 5; 2026: 5.

12 papers.

  1. 2024What Do Machine Learning Researchers Mean by "Reproducible"?cited by 9
  2. 2024Is Function Similarity Over-Engineered? Building a Benchmarkcited by 6
  3. 2025Tady: A Neural Disassembler without Structural Constraint Violationscited by 2
  4. 2025BuildBench: Benchmarking LLM Agents on Compiling Real-World Open-Source Softwarecited by 2
  5. 2025Zipf-Gramming: Scaling Byte N-Grams Up to Production Sized Malware Corporacited by 2
  6. 2025Binary BPE: A Family of Cross-Platform Tokenizers for Binary Analysiscited by 2
  7. 2026Cybersecurity is the True Frontier for Generative AI Success or Failurecited by 2
  8. 2025Binary-30K: A Heterogeneous Dataset for Deep Learning in Binary Analysis and Malware Detectioncited by 1
  9. 2026REFORGE: A Method for Benchmarking LLMs' Reverse Engineering Capabilities in Decompiled Binary Function Namingcited by 1
  10. 2026Identifier-Free Code Embedding Models for Scalable Searchcited by 0
  11. 2026ASSEMBLAGE-DEEPHISTORY: A Cross-Build Binary Dataset with Temporal Coveragecited by 0
  12. 2026When LLM Decompilers Recompile More and Preserve Lesscited by 0