---
title: "Causal Newton Optimization: Online Calibration with Iterative Local Linear Modeling and Newton Updates"
authors:
  - "Daigo Fujiwara"
  - "Tomonori Izumitani"
  - "Shohei Shimizu"
doi: "10.24963/ijcai.2026/710"
url: "https://ijcai.org/proceedings/2026/710"
pdf: "https://ijcai.org/proceedings/2026/0710.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "6380-6388"
track: "Main Track"
keywords:
  - "Machine Learning: Optimization"
  - "AI Ethics, Trust, Fairnes: Safety and robustness"
  - "Machine Learning: Online learning"
  - "Uncertainty in AI: Causality, structural causal models and causal inference"
---

# Causal Newton Optimization: Online Calibration with Iterative Local Linear Modeling and Newton Updates

Daigo Fujiwara, Tomonori Izumitani, Shohei Shimizu

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 6380-6388. DOI: [10.24963/ijcai.2026/710](https://doi.org/10.24963/ijcai.2026/710)

## Abstract

Optimization in industrial systems often involves calibrating from a semi-optimized state, where global exploration methods like Reinforcement Learning (RL) or Bayesian Optimization (BO) are inefficient or unsafe. We propose Causal Newton Optimization (CNO), an online algorithm that iteratively calibrates inputs under a known causal graph but unknown structural equations. CNO estimates local linear causal effects via additive interventions and employs a log-linear variance regression to robustly guide Newton-based updates. Evaluations on synthetic systems and a chemical plant simulator demonstrate that CNO achieves the best balance between objective improvement and robustness. While traditional PID control suits standard dynamical systems, CNO significantly outperforms RL and BO in complex structural causal models, providing the robust stability vital for safety-critical real-world applications.