Research and Publications

Current Research (Industry)

1. Multi-Agent AI Research Pipeline — Bayer Crop Science R&D, 2026–present

• Architected and deployed a production multi-agent research pipeline using LangGraph and Google ADK, consisting of four specialized agents: literature review, internal knowledge retrieval, cloud data querying, and automated model training.
• Designed the end-to-end agentic workflow to autonomously connect external genomics findings with internal domain knowledge, query relevant datasets from cloud infrastructure, and conduct model training to accelerate the research cycle.

2. Machine Learning for Large-Scale Breeding Analytics — Bayer Crop Science R&D, 2023–present

• Led ML modeling on million-scale datasets, applying CNN architectures to build end-to-end breeding analytics pipelines delivering actionable genetic improvement recommendations.
• Designed an active learning framework for high-value phenotype data acquisition in crop disease resistance; built ensemble learning models for soybean maturity prediction.

3. ML Tooling & Infrastructure — Bayer Crop Science R&D, 2023–2024

• Developed and maintained a Python ML library standardizing data processing, feature engineering, and model training workflows across the team.
• Designed scalable data pipelines for querying, transforming, and validating large-scale cloud-hosted datasets for downstream AI/ML applications.

PhD Research

The objective of my PhD research was to propose an effective methodological framework to protect and utilize manufacturing data for quality assurance from 3 perspectives: data quality, design quality, and process quality.

Data quality assurance: A blockchain-enabled approach for cyber-physical security protection in manufacturing

• Apply blockchain for important file (G-code)/data (stream data) storage using Python. Via mismatch of hash value in blocks, malicious modification on files could be detected timely and accurately.
• Utilize asymmetry encryption method to encrypt files to ciphertext; design a camouflage method to camouflage ciphertext to “data” having similar format with original one, which could reduce risk of unauthorized access on files.

Design quality assurance: Hybrid data-driven feature extraction-enabled surface modeling for process design

• Propose a hybrid feature extraction framework which consists of machine learning-based feature and statistics feature.
• Develop a robust convolutional autoencoder to extract low-dimensional features from 3D printing surface with large fraction of outliers, porosities, and shifts.

Process quality assurance: Develop machine learning-based methods to enhance the in-situ monitoring performance

• Apply LSTM-autoencoder to extract features from vibration signal for both supervised and unsupervised process monitoring.
• Propose a data augmentation method based on Augmented Time Regularized Generative Adversarial Network (ATR-GAN) to generate high-quality anomaly data for classifier training.
• Incorporate knowledge distillation framework in decentralized systems to improve neural network accuracy and training efficiency while preserving data privacy.

Journal Publications – Published

1. Shi, Z., Li, Y., & Liu, C*. (2025). Knowledge Distillation-based Information Sharing for Online Process Monitoring in Decentralized Manufacturing System, Journal of Intelligent Manufacturing, 36(3). [link]
2. Li, Y., Xie, T., Liu, C., & Shi, Z.*. (2024). Pseudo Replay-based Class Continual Learning for Online New Category Anomaly Detection in Additive Manufacturing, IISE Transactions. [link]
3. Shi, Z., Oskolkov, B., Tian, W., Kan, C., & Liu, C*. (2024). Sensor Data Protection through Integration of Blockchain and Camouflaged Encryption in Cyber-physical Manufacturing Systems. Journal of Computing and Information Science in Engineering, 24(7). [link]
4. Li, Y., Shi, Z., & Liu, C*. (2023). Transformer-enabled Generative Adversarial Imputation Network with Selective Generation (SGT-GAIN) for Missing Region Imputation. IISE Transactions, 56(9), 975–987. [link]
5. Xiao, P., Shi, Z., Liu, C., & Hagen, D*. (2023). Characteristics of Circulating Small Non-Coding RNAs in Plasma and Serum during Human Aging. Aging Medicine, 6, 35–48. [link]
6. Shi, Z., Mandal, S., Harimkar, S., & Liu, C*. (2022). Hybrid data-driven feature extraction-enabled surface modeling for metal additive manufacturing. The International Journal of Advanced Manufacturing Technology, 121(7), 4643–4662. [link]
7. Shi, Z., Mamun, A. A., Kan, C., Tian, W., & Liu, C*. (2023). An LSTM-autoencoder based online side channel monitoring approach for cyber-physical attack detection in additive manufacturing. Journal of Intelligent Manufacturing, 34, 1815–1831. [link]
8. Shi, Z., Kan, C., Tian, W., & Liu, C*. (2021). A Blockchain-based G-code protection approach for cyber-physical security in additive manufacturing. Journal of Computing and Information Science in Engineering, 21(4). [link]
9. Li, Y., Shi, Z., Liu, C.*, Tian, W., Kong, Z., & Williams, C. B. (2021). Augmented Time Regularized Generative Adversarial Network (ATR-GAN) for Data Augmentation in Online Process Anomaly Detection. IEEE Transactions on Automation Science and Engineering, 19(4), 3338–3355 (Finalist of the Data Challenge Award, QSR Section, INFORMS, 2019). [link]
10. Yu, S., Shi, Z., Aoun, M., Wu, Y., Fang, T., Fontanier, C., & Xiang, M*. (2025). Development of KASP markers and genomic prediction for winter hardiness in African Bermudagrass. Grass Research (accepted).

Journal Publications – Submitted and In Preparation

1. Wang, Z., Li, Y., & Shi, Z.*. (2026). Mutual knowledge sharing for enhancement of bearing anomaly detection in manufacturing process. IISE Transactions (under 1st round review).
2. Shi, Z., Wang, Z., & Li, Y*. (2026). Unsupervised Identification and Replay-based Detection (UIRD) for New Category Anomaly Detection in ECG Signal. Computers in Biology and Medicine (under 1st round review).
3. Oskolkov, B., Shi, Z., Tian, W., & Liu, C*. (2026). Knowledge distillation-empowered domain incremental learning with flexible model selection for smart manufacturing applications. Journal of Manufacturing Systems (under 2nd round review).
4. Xiao, P*., Shi, Z., Ma, S., Ran, J., Guo, A., & Chen, F. (2026). A dataset of comprehensive small non-coding RNA associated with reproductive behavior in poultry. To be submitted to Journal of Integrative Agriculture.

Conference Publications

1. Shi, Z., Li, Y., & Liu, C*. (2022). Knowledge Distillation-enabled Multi-stage Incremental Learning for Online Process Monitoring in Advanced Manufacturing. IEEE International Conference on Data Mining (ICDM) IncrLearn Workshop. [link]
2. Shi, Z., Liu, C*., Kan, C., Tian, W., & Chen, Y. (2021). A Blockchain-Enabled Approach for Online Stream Sensor Data Protection in Cyber-Physical Manufacturing Systems. ASME IDETC-CIE (Vol. 85376, p. V002T02A035). [link]
3. Shi, Z., Mandal, S., Harimkar, S., & Liu, C*. (2021). Surface Morphology Analysis Using Convolutional Autoencoder in Additive Manufacturing with Laser Engineered Net Shaping. Procedia Manufacturing, 53, 16–23. [link]