{
  "abstract": "Deep learning (DL) models are increasingly being deployed into radiation-challenging environments in spacecraft, particle accelerators, and, soon, space-based data centers. Instantiated as a learned orchestration of numbers stored in persistent memory, models are susceptible to continual degradation as high-energy particles corrupt data. At inference time, new data passed through the network fundamentally break the off-line assumptions and guarantees of a static model determined during training. To preserve the quality and efficacy of the models, fault tolerance against bit flips is a clear objective. In this work we developed a runtime method for repairing the quantized weights of neural networks (NNs) after bit flips on commodity, commercial off-the-shelf (COTS) hardware, without storing extra replicas. Our work extends the well-established Fault-Aware Training (FAT) paradigm by adding an alignment penalty to concentrate undamaged activations and weights into a low-rank subspace, and we then reconstruct flagged corrupted coordinates from the uncorrupted survivors. We call this method {\\it Projection Healing} and apply it to six regression benchmarks on quantized NNs. Holding a FAT checkpoint fixed and toggling only the runtime heal gives a further ${\\sim}23\\%$ reduction in the native-unit excess integral. Across those datasets the method extends the operational scrub interval by a median of $2.04\\times$ at a $+5\\%$ RMSE degradation and $1.96\\times$ at $+50\\%$, compared with $1.62\\times$ and $1.07\\times$ for classical Single Error Correction, Double Error Detection (SECDED). This introduces a new path towards designing resilient data-driven models, complementing established hardware redundancy, ECC, and FAT-based techniques.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.27621",
  "author_total": 2,
  "authors": [
    {
      "name": "Peter Xiangyuan Ma"
    },
    {
      "name": "Joshua S. Bloom"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/ddegcqq5mpbcf.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.27621",
  "fulltext_url": null,
  "id": "ddegcqq5mpbcf",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.27621",
  "language": null,
  "licence": "CC0-1.0",
  "licence_label": "CC0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/publicdomain/zero/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "astro-ph.IM"
    ],
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "hep-ex"
    ],
    [
      "arxiv",
      "nucl-ex"
    ],
    [
      "arxiv",
      "physics.ins-det"
    ],
    [
      "openalex",
      "2208|Radiation Effects in Electronics"
    ]
  ],
  "openalex": "W7223093463",
  "pdf_url": "https://arxiv.org/pdf/2610.27621",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-20",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/ddegcqq5mpbcf.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:34:04Z",
      "native_id": "2610.27621",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.27621"
    }
  ],
  "subjects": [
    {
      "field": "astronomy",
      "origin": "arxiv:astro-ph.IM",
      "subfield": "instrumentation-methods"
    },
    {
      "field": "physics",
      "origin": "arxiv:physics.ins-det",
      "subfield": "applied-instrumentation"
    },
    {
      "field": "physics",
      "origin": "arxiv:hep-ex",
      "subfield": "high-energy"
    },
    {
      "field": "physics",
      "origin": "arxiv:nucl-ex",
      "subfield": "nuclear"
    }
  ],
  "title": "A Radiation-Hardened, Fault-Aware Quantized Neural Network That Self-Heals",
  "title_text": "A Radiation-Hardened, Fault-Aware Quantized Neural Network That Self-Heals",
  "type": "preprint",
  "updated_at": "2026-10-30T12:34:04Z",
  "url": "https://research.draconitic.com/paper/ddegcqq5mpbcf",
  "venue": "arXiv",
  "versions": []
}