智慧型科技探勘-以工業4.0應用研究為例

Intelligent Technology Mining– An Industry 4.0 Application Study

何玄(Govindarajan, Usharani Hareesh)

指導教授 ： 張瑞芬 ； 張力元

國立清華大學/其它/跨院國際博士班學位學程/博士(2018年)

摘要

科技的發展對產業及經濟的進步有著舉足輕重的地位，而科技研發往往產出大量的技術文件及數據資料。科技進步和傳播透過大量的文集，如國際期刊、專利文件、開放型知識文件庫及重要國際標準等技術文件，才得以讓技術擴散他人分享，並繼續尋求精進的方法。企業若想藉著創新的科技發展新產品必須運用科技探勘找尋新科技組件。主題生成模型是強大的自然語言處理 Natural Language Processing (NLP)工具，此模型能夠分析文集及其相關關鍵字詞分布。本論文發展的技術探勘方法，稱之為 Excessive Topic Generation (ETG)，作為主題分析及視覺化的預處理框架。本論文發展出的ETG傳承了Latent Dirichlet Allocation (LDA) 的主題生成模式且具有生成字詞之間距離關係的學習能力。本研究以工業4.0之工業沉浸式科技 Industry Immersive Technology (IIT) 的科技組件探勘為案例，以系統化 ETG、LDA等方法，探討 IIT之科技發展趨勢。本研究檢索出超過11,000份技術文檔做為工業4. 0科技探勘文件庫。本研究亦以 IIT專利之國際專利分類 International Patent Classification (IPC)及合作專利分類 Cooperation Patent Classification (CPC)自動分類對 ETG的結果進行驗證及比較。智慧科技探勘使用 ETG前處理將使得集體智慧運用於工程諮詢及商業智能得以實現。

關鍵字

智慧技術組件探勘 ； 集體智慧 ； 過度主題生成 ； 工業4.0 應用 ； 工業沉浸式科技

並列摘要

Science and technology play a major role in advancing industry with significant effects on the world economy. An inevitable consequence of the technology-driven economy has led to an increase in technical document data. Technology components are now spread across a huge corpus of international publications, patents, open source repositories, blogs, and essential standards. Businesses wanting to adopt technology must conduct technology mining for key component integrations. This dissertation contributes to an advanced technology mining method named Excessive Topic Generation (ETG) as a preprocessing framework for topic term (referred to as key term) generation, visualization, and analysis. The presented ETG method inherits the topic generation characteristics from Latent Dirichlet Allocation (LDA) with a further capability to generate word distance relationships among key terms. The framework is applied to a study of the advanced manufacturing domain encapsulated through Industry 4.0. Industry 4.0 outline combines conventional manufacturing practices with increased technological outreach over a corpus of more than 11,000 technical documents across public databases. A systematic drill down for Industrial Immersive Technology (IIT) over 2,164 technical documents covering Virtual Reality (VR), Augmented Reality (AR), and Brain Machine Interface (BMI) concepts are demonstrated. A validation comparison of 741 global IIT patents against international invention indexation standards International Patent Classification (IPC) and the Cooperative Patent Classification (CPC) are conducted to validate the superior ETG results. Further, a conversational chatbot is integrated to distribute generated transformation. Intelligent technology mining using ETG enables collective intelligence for both engineering consultation and business intelligence.

並列關鍵字

Intelligence technology component mining ； collective intelligence ； excessive topic generation ； industry 4.0 application ； industrial immersive technologies

參考文獻

42. Govindarajan, U. H., Trappey, A. J., & Trappey, C. V. (2018). Immersive technology for human-centric cyber physical systems in complex manufacturing processes: a comprehensive overview of the global patent profile using collective intelligence. Complexity, 148, https://doi.org/10.1155/2018/4283634 (online published).

1. Agostini, L., Caviggioli, F., Filippini, R., & Nosella, A. (2015). Does patenting influence SME sales performance? A quantity and quality analysis of patents in Northern Italy. European Journal of Innovation Management, 18(2), 238-257.

2. Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., & Ayyash, M. (2015). Internet of things: A survey on enabling technologies, protocols, and applications. IEEE Communications Surveys & Tutorials, 17(4), 2347-2376.

3. Altman, N., & Krzywinski, M. (2017). Points of significance: clustering. Nature Methods, 14, 545-546.

4. Aristodemou, L., & Tietze, F. (2018). The state-of-the-art on Intellectual Property Analytics (IPA): A literature review on artificial intelligence, machine learning and deep learning methods for analyzing intellectual property (IP) data. World Patent Information, 55, 37-51.