Causal Subgraphs for Reconciling Experiment and First-Principles Prediction

Which linked observations justify preferring a thermodynamic interpretation over a metastable one? This review answers by treating model-experiment reconciliation as a property of a sociotechnical workflow rather than a feature that can be read from average accuracy. The focal setting is high-pressure crystallography, where a plausible conclusion may depend on an unrecorded stress state, preprocessing choice, or alternative structural interpretation. Evidence from the assigned publications is synthesized with foundational studies of calibration, distribution shift, causal structure, and responsible deployment. Four requirements follow: preserve the lineage of instrument traces, diffraction and spectroscopy products, simulation outputs, laboratory records, and analysis repositories; measure stability across relevant perturbations; connect confidence to a specific action; and maintain a route for human challenge and correction. The framework distinguishes descriptive performance from decision utility and separates uncertainty about the world from uncertainty created by the model and its evaluator. It also shows why faster inference or richer reasoning is valuable only when it improves a defined decision under a transparent resource budget. The article is a literature review and research agenda, not a report of a newly completed trial.

1. Ma, Mingjun, et al. "MuSK: Multi-Scale Knowledge Learning for Provenance-Graph Anomaly Detection." *Computer Networks* 289 (2026): 112728.
2. Li, Yuanhao, et al. "BoostAPR: Boosting Automated Program Repair via Execution-Grounded Reinforcement Learning with Dual Reward Models." *arXiv preprint arXiv:2605.09134* (2026).
3. Chen, Huawei, et al. "Crystal Structure of CaSiO3 Perovskite at 28-62 GPa and 300 K under Quasi-Hydrostatic Stress Conditions." *American Mineralogist* 103.3 (2018): 462-468.
4. King, Samuel T., and Peter M. Chen. "Backtracking Intrusions." *Proceedings of the Nineteenth ACM Symposium on Operating Systems Principles*, 2003, pp. 223-236.
5. Milajerdi, Sadegh M., et al. "HOLMES: Real-Time APT Detection through Correlation of Suspicious Information Flows." *2019 IEEE Symposium on Security and Privacy*, 2019, pp. 1137-1152.
6. Han, Xueyuan, et al. "UNICORN: Runtime Provenance-Based Detector for Advanced Persistent Threats." *Network and Distributed System Security Symposium*, 2020.
7. Pasquier, Thomas, et al. "Runtime Analysis of Whole-System Provenance." *Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security*, 2017, pp. 1601-1614.
8. Schlichtkrull, Michael, et al. "Modeling Relational Data with Graph Convolutional Networks." *The Semantic Web*, Springer, 2018, pp. 593-607.
9. Hamilton, William L., Rex Ying, and Jure Leskovec. "Inductive Representation Learning on Large Graphs." *Advances in Neural Information Processing Systems*, vol. 30, 2017.
10. Velickovic, Petar, et al. "Graph Attention Networks." *International Conference on Learning Representations*, 2018.
11. Chandola, Varun, Arindam Banerjee, and Vipin Kumar. "Anomaly Detection: A Survey." *ACM Computing Surveys*, vol. 41, no. 3, 2009, article 15.
12. Sommer, Robin, and Vern Paxson. "Outside the Closed World: On Using Machine Learning for Network Intrusion Detection." *2010 IEEE Symposium on Security and Privacy*, 2010, pp. 305-316.
13. Liao, Xiaojing, et al. "Acing the IOC Game: Toward Automatic Discovery and Analysis of Open-Source Cyber Threat Intelligence." *Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security*, 2016, pp. 755-766.