We propose a neuro-symbolic Transformer-ASP framework for interpretable real-time analytics in intelligent manufacturing systems, which meets the essential requirement for high-performance yet explainable decision-making in industrial environments. The framework integrates transformer-based neural architecture with Answer Set Programming (ASP) to process multi-modal sensor data while generating human-understandable rules. The system functions via three interconnected layers: a feature extraction layer adopting hierarchical transformers to identify temporal and cross-sensor relationships, followed by a layer where latent representations are converted into probabilistic logic predicates through neuro-symbolic methods, and concluding with an actionable rule generation layer producing interpretable decision rules verified by empirical data patterns. In contrast to traditional post-hoc interpretability techniques, our method creates a two-way interaction between neural and symbolic elements, which permits the ASP layer to actively shape feature acquisition while anchoring abstract rules in empirical data patterns. The framework interfaces with industrial control systems by compiling ASP-derived rules into executable policies and dynamically prioritizing sensor inputs based on transformer attention weights. The system, running on TensorRT-accelerated hardware with an adapted Clingo solver, supports end-to-end training by means of differentiable ASP inference while simultaneously improving prediction accuracy and rule conciseness. Experimental findings indicate the framework attains real-time operation on edge devices and grants manufacturing engineers’ actionable intelligence regarding process anomalies and equipment malfunctions. This work advances the state of the art in industrial AI by unifying the representational power of deep learning with the rigor of symbolic reasoning, thereby bridging the gap between data-driven analytics and human-interpretable decision logic.
References
[1] Liang, S., Rajora, M., Liu, X., Yue, C., Zou, P. (2018) Intelligent manufacturing systems: A review. International Journal of Mechanical Engineering and Robotics Research, 7(3), 324-330.
[2] Holzinger, A., Saranti, A., Molnár, C., Biecek, P. (2020) Explainable AI methods: A brief overview. In International Workshop on Extending Explainability to Novel Domains (pp. 13-38).
[3] Wang, F., Jiang, Y., Zhang, R., Wei, A., Xie, J., Pang, X. (2025) A survey of deep anomaly detection in multivariate time series: Taxonomy, applications, and directions. Sensors (Basel, Switzerland), 25(1), 190.
[4] Vaswani, A., Shazeer, N., Parmar, N., et al. (2017) Attention is all you need. In Advances in Neural Information Processing Systems (Vol. 30).
[5] Li, Z. (2022) Extracting spatial effects from machine learning model using local interpretation method: An example of SHAP and XGBoost. Computers, Environment and Urban Systems, 96, 101845.
[6] Hitzler, P., Sarker, M. (2022) Neuro-symbolic AI = neural + logical + probabilistic AI. Neuro-Symbolic Artificial Intelligence: The State of the Art, 342, 173-194.
[7] Wu, J. (2017) Introduction to convolutional neural networks. National Key Lab for Novel Software Technology, Nanjing University, China, 5(23), 495.
[8] Yu, Y., Si, X., Hu, C., Zhang, J. (2019) A review of recurrent neural networks: LSTM cells and network architectures. Neural Computation, 31(7), 1235-1270.
[9] Parmar, N., Vaswani, A., Uszkoreit, J., et al. (2018) Image transformer. In International Conference on Machine Learning (pp. 4055-4064).
[10] Okokpujie, I. P., Bolu, C. A., Ohunakin, O. S., Akinlabi, E. T., Adelekan, D. S. (2019) A review of recent applications of machining techniques based on the phenomena of CNC machining operations. Procedia Manufacturing, 35, 1054-1060.
[11] Xiao, Y., Zheng, S., Shi, J., Du, X., Hong, J. (2023) Knowledge graph-based manufacturing process planning: A state-of-the-art review. Journal of Manufacturing Systems, 70, 417-435.
[12] Shakil, N. A. F., Mia, R., Ahmed, I. (2023) Applications of AI in cyber threat hunting for advanced persistent threats (APTs): Structured, unstructured, and situational approaches. Journal of Applied Big Data Analytics, Decision-Making, and Predictive Modelling Systems, 7(12), 19-36.
[13] Zhou, T., Sheng, H., Howley, I. (2020) Assessing post-hoc explainability of the BKT algorithm. Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (pp. 407–413).
[14] Zimmer, M., Feng, X., Glanois, C., Jiang, Z., Zhang, J. (2021) Differentiable logic machines. arXiv preprint arXiv:2102.11529.
[15] Li, Y., Liu, X., Starly, B. (2024) Manufacturing service capability prediction with graph neural networks. Journal of Manufacturing Systems, 74, 291-301.
Share and Cite
Wang, Y., Wan, Y. (2024) Neuro-Symbolic Transformer-ASP Framework for Interpretable Real-Time Analytics in Intelligent Manufacturing Systems. Scientific Research Bulletin, 1(6), 51-62.