Hi! I'm Chang,

2022.jpg
CST 3-118, 111 College Pl, Syracuse, NY 13210

cliu57@syr.edu

a PhD student in the Department of Electrical Engineering and Computer Science at Syracuse University. It’s my fortune to be advised by Kristopher Micinski.

My research explores the core principles of secure and verifiable programming, with a specific focus on compilation and reverse engineering. I’m particularly interested in (what I think) the “Laplace’s demon myth” in reverse engineering: whether a deterministic (or nearly deterministic) compilation process can be reversed.

To tackle this, I’ve developed and implemented scalable systems for curating large datasets of machine executables. I’ve also benchmarked machine learning algorithms on various binary analysis tasks. My expertise extends to compilers (with CompCert and LLVM), logic programming, and the application of large language model (LLM) agents.

Publications

  1. NeurIPS
    Assemblage: Automatic Binary Dataset Construction for Machine Learning
    Chang Liu*, Rebecca Saul*, Yihao Sun, Edward Raff, Maya Fuchs, Townsend Southard Pantano, James Holt, and Kristopher Micinski
    In Proceedings of the Thirty-Eighth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track, 2024
    Code PDF
  2. NeurIPS
    Is Function Similarity Over-Engineered? Building a Benchmark
    Rebecca Saul, Chang Liu, Noah Fleischmann, Richard J Zak, Kristopher Micinski, Edward Raff, and James Holt
    In Proceedings of the Thirty-Eighth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track, 2024
    Code PDF

Services

Reviewer

2025: AAAI AICS Workshop
2026: NeurIPS Datasets & Benchmarks Track