{
  "abstract": "Materials companies must jointly develop products that meet competing requirements while keeping their formulations, processes and models private. Sample exchange alone can leave their searches poorly coordinated. We introduce Minimum Knowledge Collaboration (MKC), which connects private Bayesian optimizers through bounded candidate descriptions and company-specific search directions. Mathematical analysis links representation error, feasible private responses and information-limited search. In 256 controlled synthetic programs, all methods pursued the same reachable product target, defined as 100%. After 64 complete-product tests, MKC reached 92.4%, the information-rich reference 93.8%, and reusable better, neutral or worse feedback 6.0%. By 2,048 tests, ternary feedback reached MKC's result in 30 of 32 simplest bilateral programs; attainment across all 1,024-candidate programs was 12.5%. A cross-study scenario calculation gave a 0.0063% complete-state reconstruction index. This assumption-based design screen complements the measured reconstruction results. MKC demonstrates coordinated product development with bounded disclosure and supports industrial evaluation.",
  "abstract_shown": true,
  "affiliations": [
    "MateriaX"
  ],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Yakov Kutsovsky",
      "orcid": "0000-0002-5920-2040"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/3y33dn3zv6e5b.bib",
  "cited_by": 0,
  "doi": "10.26434/chemrxiv.15009872/v1",
  "fulltext_url": null,
  "id": "3y33dn3zv6e5b",
  "issns": [],
  "landing_url": "https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009872/v1",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "2505|Machine Learning in Materials Science"
    ]
  ],
  "openalex": "W7220429717",
  "pdf_url": "https://chemrxiv.org/doi/pdf/10.26434/chemrxiv.15009872/v1",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-05",
  "publisher": "American Chemical Society (ACS)",
  "ris": "https://research.draconitic.com/paper/3y33dn3zv6e5b.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:17:22Z",
      "native_id": "10.26434/chemrxiv.15009872",
      "source": "chemrxiv",
      "url": "https://chemrxiv.org/doi/full/10.26434/chemrxiv.15009872/v1"
    }
  ],
  "subjects": [
    {
      "field": "chemistry",
      "origin": "openalex:2505|Machine Learning in Materials Science",
      "subfield": "materials-polymers"
    }
  ],
  "title": "Minimum Knowledge Collaboration enables joint materials development through protected optimization",
  "title_text": "Minimum Knowledge Collaboration enables joint materials development through protected optimization",
  "type": "preprint",
  "updated_at": "2026-10-30T12:17:22Z",
  "url": "https://research.draconitic.com/paper/3y33dn3zv6e5b",
  "venue": "ChemRxiv",
  "versions": []
}