2026 · Preprint

Echo: Learning-based Matching Decompilation using Trusted Back Translation

Echo iteratively generates source and compiler configurations, recompiles candidates, and uses assembly differences to guide configuration search and code refinement. It evaluates matching on function benchmarks and selected functions from a Mirai binary.

Scope and limitations

Matching means equality of normalized function assembly under evaluated compilation settings, not byte-for-byte recovery of a whole executable. Tests cover selected x86 GCC/LLVM settings and 62 Mirai functions; passing finite harness tests is not a general equivalence proof. Long optimized functions and configurations outside the searched space remain limitations.

Why it fits this collection

Adds compiler-feedback-driven matching decompilation while separating its matching criterion from test-only recovery.

Publication and review details

Venue
arXiv
Status
Preprint
Publication / posting date
September 16, 2026
Date source
First arXiv submission date from the reviewed submission history; no formal publication was verified.
First arXiv submission
September 16, 2026
arXiv
2609.18706
Reviewed arXiv version
v1

Latest reviewed version is arXiv v1. The Open Science section uses the placeholder anonymous.4open.science/xxxx, so no usable author artifact, model revision or code/data license was verified. The article uses the arXiv non-exclusive distribution license. No upstream code, models or malware were downloaded or run.

Sources reviewed

Source review only; experiments were not rerun for this collection. Cite the original authors using the linked paper or BibTeX.

Primary sources

  1. https://arxiv.org/abs/2609.18706v1
  2. https://arxiv.org/html/2609.18706v1