{
  "annotation_id": "b245df51-ddc8-4219-af07-9f9239c63001",
  "date_created": "2026-03-02T18:00:43.427000Z",
  "date_modified": "2026-03-02T18:00:43.427000Z",
  "file_hash": "65bdc1b0a632acc17ff9f9dc6dcf0730e81f422fa73cd642f7da30268b72a01c",
  "private": false,
  "record": {
    "abstract": "Despite of their success, the results of first-principles quantum mechanical\ncalculations contain inherent numerical errors caused by various\napproximations. We propose here a neural-network algorithm to greatly reduce\nthese inherent errors. As a demonstration, this combined quantum mechanical\ncalculation and neural-network correction approach is applied to the evaluation\nof standard heat of formation $\\DelH$ and standard Gibbs energy of formation\n$\\DelG$ for 180 organic molecules at 298 K. A dramatic reduction of numerical\nerrors is clearly shown with systematic deviations being eliminated. For\nexamples, the root--mean--square deviation of the calculated $\\DelH$ ($\\DelG$)\nfor the 180 molecules is reduced from 21.4 (22.3) kcal$\\cdotp$mol$^{-1}$ to 3.1\n(3.3) kcal$\\cdotp$mol$^{-1}$ for B3LYP/6-311+G({\\it d,p}) and from 12.0 (12.9)\nkcal$\\cdotp$mol$^{-1}$ to 3.3 (3.4) kcal$\\cdotp$mol$^{-1}$ for\nB3LYP/6-311+G(3{\\it df},2{\\it p}) before and after the neural-network\ncorrection.",
    "arxiv_id": "physics/0306075",
    "authors": [
      "LiHong Hu",
      "XiuJun Wang",
      "LaiHo Wong",
      "GuanHua Chen"
    ],
    "categories": [
      "physics.data-an",
      "physics.chem-ph",
      "physics.comp-ph"
    ],
    "title": "Combined first--principles calculation and neural--network correction   approach as a powerful tool in computational physics and chemistry",
    "url": "https://arxiv.org/abs/physics/0306075"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "34555b51-d817-4358-a68a-6465171d6aa9",
    "id": "arXiv Dataset IDs",
    "type": "Model",
    "variant": "snapshot-2026-03-01",
    "version": "0.1.0"
  },
  "user_id": 1000002
}