本研究以台灣的多元件加工業相關產業為例，描述一些中小企業製造商遇到的困境。由於高競爭的特性，在面對眾多客戶的訂單時，難以制定統一的標準去規範客戶的零件格式，往往需要耗費大量的人力及時間進行分類及估價，這不僅增加了無形成本，也使得資料庫建置困難，進而阻礙了供應鏈整合。
為解決這些問題，本研究透過文獻回顧比較不同的資料儲存方式，提出一套以混合式區塊鏈及語言模型為基礎的輕量級平台架構，結合了區塊鏈的安全性、追溯性以及關聯式資料庫的高效率查詢能力。於此基礎上，本研究新增在此架構中新增一個資料正規化模組，透過語言模型將進入企業內部的資料流進行自動化清洗。
在研究中，主要比較了三種語言模型對無上下文的短字詞辨識的表現。實驗結果顯示，三種模型均有良好表現，其中DistilBERT兼具最快的訓練速度及最佳的性能。透過此正規化模組，能夠改善原先品質不佳的資料流，建置完整的資料庫。
結合語言模型與混合式區塊鏈的供應鏈整合平台，不僅能確保資料的透明性與安全性，還能促進上下游企業間的協作，減少人力及時間成本，提升供應鏈管理的競爭力和運作效率，對於面臨市場競爭壓力和資源限制的台灣中小企業製造商具有重要意義。

This study takes Taiwan’s multi-component processing industry as an example to describe the challenges faced by SMEs. Due to the highly competitive nature of the industry, it is difficult to standardize the component formats of various customer orders, often requiring substantial manpower and time for classification and quotation. This not only increases intangible costs but also makes database construction difficult, thereby hindering supply chain integration.
To address these issues, this study reviews the literature to compare different data storage methods and proposes a lightweight platform architecture based on hybrid blockchain and language models. This architecture combines the security and traceability of blockchain with the efficient query capabilities of relational databases. On this basis, a data normalization module is added to the architecture to automatically clean the data entering the enterprise through language models.
The study mainly compares the performance of three language models in recognizing context-free short words. Experimental results show that all three models perform well, with DistilBERT achieving the fastest training speed and best performance. This normalization module can improve the quality of previously poor data flows and build a complete database.
The integration of language models with the hybrid blockchain supply chain integration platform not only ensures data transparency and security but also promotes collaboration among upstream and downstream enterprises, reduces labor and time costs, and enhances the competitiveness and operational efficiency of supply chain management. This has significant implications for Taiwanese SMEs facing market competition pressures and resource constraints.

論文審定書 i
摘要 ii
Abstract iii
Table of Figures vi
Table of Tables viii
Chapter 1 Introduction 1
1.1 Research Background 1
1.2 Research Motivation 2
1.3 Research Purpose 7
Chapter 2 Literature Review 9
2.1 The Development of Blockchain 9
2.2 Advantages and Challenges of Blockchain in Manufacturing 12
2.3 Blockchain and Traditional Relational Databases 14
2.4 Hybrid Blockchain 17
2.5 Short Text Classification Models and Methods 20
2.6 Related Patents 23
Chapter 3 System Architecture and Method 29
3.1 Platform Architecture Design 31
3.1.1 Overview of Platform Architecture Design 31
3.1.2 Data Flow and Processing Procedures 33
3.2 Data Normalization Module 35
3.2.1 Data Preprocessing 35
3.2.2 Language Models 36
3.2.3 Training Methods 41
3.2.3 Validation Methods 44
Chapter 4 Results 46
4.1 Learning Rate 46
4.2 Fully-Connected Layer and Dropout 50
4.3 Incremental Learning 54
Chapter 5 Conclusion 58
5.1 Discussion 58
5.2 Future Research 60
Reference 62

Barman, R., Deshpande, S., Agarwal, S., Inamdar, U., Devare, M., & Patil, A. (2019). Transfer learning for small dataset. Proceedings of the National Conference on Machine Learning.
陳立武 (2020)。 中華民國專利號 202008245A。
Chen, S., Zhang, J., Shi, R., Yan, J., & Ke, Q. (2018). A comparative testing on performance of blockchain and relational database: Foundation for applying smart technology into current business systems. In Distributed, Ambient and Pervasive Interactions: Understanding Humans: 6th International Conference, DAPI 2018, Held as Part of HCI International 2018, Las Vegas, NV, USA, July 15–20, 2018, Proceedings, Part I 6 (pp. 21-34). Springer International Publishing.
邱鸿霖 (2018)。世界智慧財產權組織專利號 2018157778A1 。
褚春燕、王剑 (2020)。中國國家知識產權局專利號 111898963A。
Chowdhury, M. J. M., Colman, A., Kabir, M. A., Han, J., & Sarda, P. (2018). Blockchain versus database: A critical analysis. 2018 17th IEEE International conference on trust, security and privacy in computing and communications/12th IEEE international conference on big data science and engineering (TrustCom/BigDataSE).
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.

Gurucharan, S. B., Harshavardhan, V., Gokarnkar, S. P., & Ravishankar, B. (2020, February). Adoption of blockchain in the supply chain to improve quality of product and customer service in manufacturing sectors. In 2020 International Conference on Mainstreaming Block Chain Implementation (ICOMBI) (pp. 1-8). IEEE.
Honglin QIU (2019). World Intellectual Property Organization Patent No. 2019195071A1.
Jordan Simons (2024). United States Patent No. 20240039924A1.
Kao, A., Niraula, N. B., & Whyatt, D. (2019, June). Part name normalization. In 2019 IEEE International Conference on Prognostics and Health Management (ICPHM) (pp. 1-6). IEEE.
Kici, D., Malik, G., Cevik, M., Parikh, D., & Basar, A. (2021). A BERT-based transfer learning approach to text classification on software requirements specifications. Canadian AI.
Leng, J., Zhou, M., Zhao, J. L., Huang, Y., & Bian, Y. (2020). Blockchain security: A survey of techniques and research directions. IEEE Transactions on Services Computing, 15(4), 2490-2510.
Marinho, S. C., Costa Filho, J. S., Moreira, L. O., & Machado, J. C. (2020). Using a hybrid approach to data management in relational database and blockchain: A case study on the E-health domain. 2020 IEEE International Conference on Software Architecture Companion (ICSA-C).
Masana, M., Liu, X., Twardowski, B., Menta, M., Bagdanov, A. D., & Van De Weijer, J. (2022). Class-incremental learning: survey and performance evaluation on image classification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(5), 5513-5533.
Mukherjee, P., & Pradhan, C. (2021). Blockchain 1.0 to blockchain 4.0—The evolutionary transformation of blockchain technology. In Blockchain technology: applications and challenges (pp. 29-49). Springer.

Muzammal, M., Qu, Q., & Nasrulin, B. (2019). Renovating blockchain with distributed databases: An open source system. Future generation computer systems, 90, 105-117.
Pai, Y. (2022, August). An Information System Framework for Managing Battery Life Cycle Tracking and Reuse Based on Hybrid Blockchain and NFT-Based Smart Contract. https://hdl.handle.net/11296/nfdfad
Raja Santhi, A., & Muthuswamy, P. (2022). Influence of blockchain technology in manufacturing supply chain and logistics. Logistics, 6(1), 15.
Sanh, V., Debut, L., Chaumond, J., & Wolf, T. (2019). DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108.
Silvio Micali (2020). World Intellectual Property Organization Patent No. 2020123538A1.
Taneja, K., & Vashishtha, J. (2022). Comparison of transfer learning and traditional machine learning approach for text classification. 2022 9th International Conference on Computing for Sustainable Global Development (INDIACom).
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in neural information processing systems, 30.
Zheng, Z., Xie, S., Dai, H., Chen, X., & Wang, H. (2017). An overview of blockchain technology: Architecture, consensus, and future trends. 2017 IEEE international congress on big data (BigData congress).
Zhou, D.-W., Wang, Q.-W., Qi, Z.-H., Ye, H.-J., Zhan, D.-C., & Liu, Z. (2023). Deep class-incremental learning: A survey. arXiv preprint arXiv:2302.03648.
Gurucharan, S., et al. (2020). Adoption of blockchain in the supply chain to improve quality of product and customer service in manufacturing sectors. 2020 International Conference on Mainstreaming Block Chain Implementation (ICOMBI), IEEE.