{
  "added_at": "2026-10-20",
  "arxiv_id": "2610.23546",
  "arxiv_version_reviewed": "v1",
  "authors": [
    "Junlin Li",
    "Jianjun Huang",
    "Wei You",
    "Bin Liang"
  ],
  "bibtex_url": "https://awesomerea.org/papers/citations/paper-bscout.bib",
  "code_url": null,
  "date_basis": "First arXiv submission; the reviewed version is v1, not a verified proceedings publication.",
  "date_basis_zh": "采用 arXiv 首次提交日期；审阅版本为 v1，并非已经核实的会议录出版。",
  "date_precision": "day",
  "description_en": "BScout matches decompiled functions to a source corpus using chunked embeddings and inlining-aware training. Existing matches provide positional clues to expand candidates; LLM selection and structural filters refine the resulting correspondences.",
  "description_zh": "BScout 通过分块嵌入及考虑内联的训练，将反编译函数匹配到源码语料；已有匹配提供位置线索以扩充候选，再由大语言模型选择及结构过滤改进对应关系。",
  "doi": null,
  "entry_url": "https://awesomerea.org/papers/paper-bscout/",
  "entry_url_zh": "https://awesomerea.org/zh/papers/paper-bscout/",
  "evidence_urls": [
    "https://arxiv.org/abs/2610.23546",
    "https://arxiv.org/html/2610.23546v1"
  ],
  "first_arxiv_date": "2026-10-16",
  "id": "paper-bscout",
  "json_url": "https://awesomerea.org/papers/paper-bscout/record.json",
  "kind": "research",
  "markdown_url": "https://awesomerea.org/papers/paper-bscout/index.md",
  "markdown_url_zh": "https://awesomerea.org/zh/papers/paper-bscout/index.md",
  "metadata_notes": "Reviewed arXiv v1 and its four-author order. The paper promises a replication package upon formal peer-reviewed publication; no dedicated public implementation or license was verified.",
  "metadata_notes_zh": "审阅 arXiv v1 并采用其四位作者顺序。论文承诺在同行评审版本正式出版后发布复现包，未核实专用公开实现或许可。",
  "pdf_url": "https://arxiv.org/pdf/2610.23546v1",
  "project_url": null,
  "publication_date": "2026-10-16",
  "publication_status": "preprint",
  "record_type": "paper",
  "runtime_tested": false,
  "scope_limitations_en": "BinKit-BCSA covers 3,904 small-utility binaries; Large-10 covers larger projects only at x86-64/O2. Some anomaly thresholds come from FreeCAD in Large-10. Incomplete ground truth, sparse anchors and source-corpus coverage limit conclusions. Function chunks are encoded independently, and extra refinement trades precision and cost against recall. This is correspondence retrieval, not proof of source recovery or semantic equivalence.",
  "scope_limitations_zh": "BinKit-BCSA 包含 3,904 个小型工具二进制，Large-10 的较大项目仅采用 x86-64/O2 构建；部分异常阈值来自 Large-10 中的 FreeCAD。不完整真值、稀疏锚点及源码语料覆盖限制结论。函数分块独立编码，增加改进轮次会在召回、精度和成本之间权衡。这是对应关系检索，不是源码恢复或语义等价证明。",
  "targets": [
    "binary"
  ],
  "title": "Recall-Oriented Precise Binary-to-Source Function Matching",
  "title_zh": "面向召回率的精确二进制到源码函数匹配",
  "url": "https://arxiv.org/abs/2610.23546",
  "venue": "arXiv",
  "verification": "source-reviewed",
  "verified_at": "2026-10-20",
  "why_fits_en": "Helps identify source origins and reused components in stripped binaries, including inlined functions.",
  "why_fits_zh": "帮助识别去符号二进制中的源码来源及复用组件，并考虑内联函数。",
  "year": 2026
}