Keywords
multivariate time series
causal inference
fine-grained anomaly detection
Abstract
Semiconductor manufacturing involves hundreds of process variables from deposition, etching, lithography, cleaning, and thermal control equipment. Abnormal wafer quality is often caused by small deviations in a few upstream variables, but these deviations may propagate through multiple process stages and appear as complex multivariate time series anomalies. This study proposes a causal variable attribution method for fine-grained anomaly detection in semiconductor process time series. The method first constructs a process-level causal graph using conditional independence testing and time-lagged Granger constraints. A counterfactual reconstruction module is then used to estimate expected sensor behavior under normal causal conditions. Variable-level anomaly scores are calculated by comparing observed signals with counterfactual trajectories, allowing the model to distinguish root-cause variables from downstream affected variables. Experiments are conducted on a semiconductor process dataset containing 186 production tools, 412 process variables, and 15-second records collected across 138 production days. The dataset includes 96 million timestamped observations and 2,240 engineer-confirmed abnormal process segments, including chamber pressure drift, gas-flow instability, abnormal RF power fluctuation, wafer temperature deviation, and endpoint-detection failure. The proposed method reduces median root-cause localization time from 38.5 minutes to 11.7 minutes compared with a correlation-based temporal baseline. The false alarm volume is controlled at 2.6 cases per production tool per week. The mean reciprocal rank for root-cause variable localization reaches 0.842, and 1,780 abnormal segments are assigned to interpretable causal paths. The model completes one full production-line assessment in 8.9 minutes. These results show that causal variable attribution can improve fine-grained anomaly detection and root-cause analysis in high-dimensional semiconductor process monitoring.
References
1. Zhang, Z., Tong, Y., & Gao, Y. (2026). Retrieval-Augmented Generation with Low-Latency Deployment for Vertical Domains Question Answering: A Case Study on Economic Resource Platforms.
2. Tarcsay, B. L., Bárkányi, Á., Chován, T., & Németh, S. (2022). A dynamic principal component analysis and fréchet-distance-based algorithm for fault detection and isolation in industrial processes. Processes, 10(11), 2409.
3. Qi, C., & Qiao, X. (2026). Using AI to Monitor AI: Automated Operations Through Log-Driven Intelligence. Available at SSRN 6795840.
4. Ahmadi, H., Mahdimahalleh, S. E., Farahat, A., & Saffari, B. (2025). Unsupervised time-series signal analysis with autoencoders and vision transformers: A review of architectures and applications. arXiv preprint arXiv:2504.16972.
5. Chen, X., Xiao, H., Zeng, Z., Zhang, S., & Xiao, R. (2025). Fine-Grained Multivariate Time Series Anomaly Detection via Causal Inference. Knowledge-Based Systems, 114765.
6. Winslow, G. (2025). Graph-Enhanced Temporal Modeling for Long-Sequence Forecasting: Dynamic Dependency Learning and Multi-Scale Feature Fusion. Applied Artificial Intelligence and Computing Systems, 1(1).
7. Koistinen, K., Hellsten, K., Herttuainen, J., & Kaski, K. K. (2026). Spatio-Temporal Attention Graph Neural Network: Explaining causalities with Attention. IEEE Access.
8. Su, D., & Dong, Y. (2026). Classroom-Based Assessment with Bayesian Learning Analytics for Instructional Decision-Making in ASD Inclusive Education.
9. Xu, T., Zhu, W., & Zhang, J. (2026). Stability and Consistency of Explainable Deep Learning Methods in Credit Risk Assessment. Available at SSRN 6893938.
10. Abshari, D., Shi, P., Fu, C., Sridhar, M., & Du, X. (2024). INVARLLM: LLM-assisted Physical Invariant Extraction for Cyber-Physical Systems Anomaly Detection. arXiv preprint arXiv:2411.10918.
11. Zhou, Y., & Long, L. (2026). Causal Effect Evaluation of Personalized Reminder Strategies on Government Welfare Program Enrollment: A Propensity Score Matching Approach. Journal of Computing Innovations and Applications, 4(1), 106-116.
12. Ma, Y. (2026). Industrial Financial Risk Prediction Model Based on Graph Neural Network and Knowledge Graph Inference. Journal of Circuits, Systems and Computers, 35(16), 2650103.
13. Gui, H., Wang, B., Lu, Y., & Fu, Y. (2025). Computational Modeling-Based Estimation of Residual Stress and Fatigue Life of Medical Welded Structures.
14. Zamanzadeh Darban, Z., Webb, G. I., Pan, S., Aggarwal, C., & Salehi, M. (2024). Deep learning for time series anomaly detection: A survey. ACM Computing Surveys, 57(1), 1-42.
15. Reiter, T., & Schoedel, R. (2024). Never miss a beep: Using mobile sensing to investigate (non-) compliance in experience sampling studies. Behavior Research Methods, 56(4), 4038-4060.
16. Chen, F., Liang, H., Li, S., Yue, L., & Xu, P. (2025). Design of Domestic Chip Scheduling Architecture for Smart Grid Based on Edge Collaboration.
17. Liang, R., Ye, Z., Liang, Y., & Li, S. (2025). Deep Learning-Based Player Behavior Modeling and Game Interaction System Optimization Research.
18. Suzuki, K., Yamamichi, M., Osada, Y., Ushio, M., Nakajima, K., Masuya, H., & Shimizu, S. (2026). Advances in causal discovery methods for ecological time series. Biological Reviews.
19. Yang, M., Wu, J., Tong, L., & Shi, J. (2025). Design of Advertisement Creative Optimization and Performance Enhancement System Based on Multimodal Deep Learning.
20. Yin, J., Rao, H., & Huang, Y. (2026, March). Dynamic Modeling and Heterogeneity Analysis of Platform User Behavior Time Series. In 2026 International Conference on AI in Education Technology and Applications (AIETA) (pp. 30-33). IEEE.
21. Yalim, C., Unal, R., & Handley, H. A. (2026). A Three‐Stage Causal Root‐Cause Diagnostic Protocol for Nonstationary Industrial Time Series Data. International Journal of Intelligent Systems, 2026(1), 3436744.
22. Yang, J. (2026). Stage‐Coupled Computational Framework for Stratified Accessibility and Equity Analysis in Community‐Based Elderly Care Services.
23. Zhang, Z. (2026). A Study on Return Optimization in E-commerce for Complex Consumer Goods Driven by Installation Information Quality. Available at SSRN 6734720.
24. Rozema, L. A., Strömberg, T., Cao, H., Guo, Y., Liu, B. H., & Walther, P. (2024). Experimental aspects of indefinite causal order in quantum mechanics. Nature Reviews Physics, 6(8), 483-499.
25. Gao, G., Gao, R., Lu, C., Gao, R., & Kuang, Y. (2026, March). Security Governance Methods and Quantitative Evaluation for Enterprise SMS and Verification Code Systems. In 2026 International Conference on Generative Artificial Intelligence and Information Security (GAIIS) (pp. 455-458). IEEE.
26. Li, Y., & Liu, S. (2026, May). A Study on Dynamic Optimization of Alerting Policies and Multi-Agent Decision-Making Mechanisms in Cloud Environments. In 2026 7th International Seminar on Artificial Intelligence, Networking and Information Technology (AINIT) (pp. 703-706). IEEE.
27. Kraus, C., Kautzky, A., Watzal, V., Gramser, A., Kadriu, B., Deng, Z. D., ... & Kasper, S. (2023). Body mass index and clinical outcomes in individuals with major depressive disorder: Findings from the GSRD European Multicenter Database. Journal of affective disorders, 335, 349-357.
28. Xu, T., Zhang, J., & Zhu, W. (2026). Reproducible Modeling Pipelines and Cross-Window Stability in Subprime Auto Loan Credit Risk Assessment. Available at SSRN 6893861.
29. Pei, Z., Huang, Q., & Wang, S. (2026). When LLMs Develop Languages: Symbolic Communication for Efficient Multi-Agent Reasoning. arXiv preprint arXiv:2606.29354.
30. Seid Ahmed, Y., Abubakar, A. A., Arif, A. F. M., & Al-Badour, F. A. (2025). Advances in fault detection techniques for automated manufacturing systems in industry 4.0. Frontiers in Mechanical Engineering, 11, 1564846.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Copyright (c) 2026 Jinwoo Kim, Hyejin Park (Author)