{
  "abstract": "Chemical process networks combine partially observed dynamics, reaction and phase behavior,\n\ntransport, material history, and decisions across unequal timescales. This paper develops a\n\nmathematical specification for integrating these features with learned models and coordinated\n\noptimization in the Astralithos framework, using Lithos as a development and verification platform.\n\nThe formulation distinguishes physical states, controls, uncertain parameters, modes,\n\nmeasurements, and economic objectives. Established balances, differential-algebraic models,\n\nexact-lumping conditions, inference methods, and optimization results are connected through\n\nexplicit conditions for dynamic closure, feasible lifting, error propagation, and decision consistency.\n\nSelf-contained derivations state the assumptions needed for these connections. Measurement\n\nuncertainty, process variability, model discrepancy, and numerical error remain distinct.\n\nThree reproducible synthetic studies examine the specification. A historical three-stage\n\noptimization example is reproduced with a feasible sequential comparator. A constructed\n\nsufficient representation is compared with raw coordinates on the same physical problem and\n\nevaluation budget. A coupled reaction, filtration, and drying model illustrates state memory and\n\nevolving conditional response surfaces. These studies establish case-specific numerical results;\n\nthey do not establish a general low-dimensional representation, real-time industrial performance, or\n\nadded value from a learned reasoning model. The resulting framework supplies explicit\n\nmathematical and evidential conditions for testing those claims.",
  "abstract_shown": true,
  "affiliations": [
    "Astralithos Inc"
  ],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Palash Gangopadhyay"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/2vsy0krckh2zf.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23600060",
  "fulltext_url": null,
  "id": "2vsy0krckh2zf",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23600060",
  "language": "eng",
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [],
  "openalex": null,
  "pdf_url": "https://zenodo.org/api/records/23600060/files/Astralithos_Multidimensional_AI_Mathematical_Foundations_v0.3.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-23",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/2vsy0krckh2zf.ris",
  "sources": [
    {
      "first_seen": "2026-11-06T12:21:05Z",
      "native_id": "23600060",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23600060"
    }
  ],
  "subjects": [],
  "title": "Mathematical foundations of multidimensional AI for dynamic chemical process networks",
  "title_text": "Mathematical foundations of multidimensional AI for dynamic chemical process networks",
  "type": "preprint",
  "updated_at": "2026-11-06T12:21:05Z",
  "url": "https://research.draconitic.com/paper/2vsy0krckh2zf",
  "venue": "Zenodo",
  "versions": []
}