Cite

If you use this software, please cite it. Every release is archived on Zenodo, and GitHub's "Cite this repository" reads the same metadata from CITATION.cff.

@software{gradina_causal_hybrid_control,
  author  = {Gradina, Ilia},
  title   = {causal-hybrid-control: physics-structured dynamics with a learned causal residual},
  year    = {2026},
  version = {0.18.0},
  doi     = {10.5281/zenodo.21737789},
  license = {Apache-2.0},
  url     = {https://github.com/causaldyn/causal-hybrid-control}
}

The doi is the concept DOI: it resolves to the newest release rather than freezing at the version above, so a reader following the citation lands on current code.

CITATION.cff

cff-version: 1.2.0
title: "causal-hybrid-control"
message: "If you use this software, please cite it using these metadata."
type: software
authors:
  - given-names: Ilia
    family-names: Gradina
    email: ilia.gradina@gmail.com
    orcid: "https://orcid.org/0009-0008-0911-8765"
repository-code: "https://github.com/causaldyn/causal-hybrid-control"
# the concept DOI, which resolves to the newest release rather than to one version
doi: "10.5281/zenodo.21737789"
license: Apache-2.0
version: 0.18.0
date-released: "2026-10-13"
abstract: >-
  Physics-structured dynamics with a learned causal residual, controlled by constrained optimal
  control and MPC, and hedged on offline confounded data by an explicit pessimism and support
  layer. Includes causal identification of a control-affine residual's control channel from
  confounded logs, sensitivity-aware control under hidden confounding, a pointwise control-barrier
  audit under a partially identified effect, and results at the causal/control seam whose algebraic
  cores are machine-checked in Rocq.
keywords:
  - causal inference
  - optimal control
  - model predictive control
  - system identification
  - scientific machine learning
  - hybrid models
  - offline reinforcement learning
  - jax

The papers that develop the results behind the library are not public yet; this page will list them once they are.