Jump to the content zone at the center

MediaTek Research, Taipei City Government, NTUT Join Forces to Explore New Frontier of Chinese Large-scale Generative AI

/001/Upload/330/relpic/31011/9078238/d87450d9-6927-4b94-91ff-7fd213ce78a4.jpg

MediaTek Research, the unit responsible for research on artificial intelligence under the Global MediaTek Group, signed a memorandum of understanding with Taipei City Government Department of Information Technology (DOIT) and National Taipei University of Technology (NTUT) on November 23.  Under the MOU, MediaTek licenses its independently-developed, large-scale Mandarin language model to DOIT, which is working with NTUT on the deployment and applications of the technology. By introducing generative artificial intelligence (AI), we seek to create a smart working style for city employees characterized by strengthened cybersecurity and high productivity. This is also an important milestone in promoting the development of Mandarin large-scale generative AI, made possible by a partnership among the industry, the academia, and the public sector. 


Dr. Shiu Da-shan, the managing director of MediaTek Research, noted: “As generative AI promises to become a new and powerful tool in enhancing production, MediaTek Research seeks to speed up the process of delivering this tool to those who needs it by tapping upon our capacity in research and development.”  NTUT views the partnership among the three entities as highly important, with vice president Thomas Yang attending the signing ceremony in person on behalf of the university. During his address, Yang pointed out that: “In recent years, NTUT has not limited its capacity solely to the research and development of industry hardware. It is proactively exploring the fields of digital transition and generative AI software. As a member of the academia, NTUT is collaborating with Taipei City Government and MediaTek Research on this project. It seeks to contribute its research capacity to assist in areas such as optimizing prompts, upgrading office automation, improving public services, and diversifying smart city applications. The development of generative AI through academic innovation has the potential of producing results which may exceed industry expectations, and we look forward to numerous collaboration cases to bring about more successful developments in the field of digital transition.”


With Taipei City serving as the project’s living labs, Commissioner Chao Shih-lung remarked: “In addition to serving the public, Taipei City Government also places an emphasis on the management of internal work-related knowledge. At the current stage, city employees who need to carry out their work by searching for data and regulations across different platforms will have to invest significant time and effort to familiarize themselves with the workflow and contents of the respective interfaces. To help city employees reduce the work burden, we seek to introduce generative AI as a tool to facilitate such processes, allowing users to look up information on related regulations or rulings through natural language inquiries. DOIT also utilized the “Guideline on Generative AI Utilization for Executive Yuan and Subsidiary Agencies/Institutions”  as reference, hoping that city employees can take advantage of innovative generative AI tools safely to help them improve administrative efficiency.”


The three-party collaboration will be carried out under the Proof-of-Concept framework, building upon the foundation of MediaTek Research’s large-scale Chinese language model to develop a data-searching chatbot that will assist city employees in raising work productivity. This will go a long way to replace the traditional, time-consuming approach for document searches and verbal inquiries by taking full advantage of AI assistance. The model adopts on-premise licensing, allowing the data to remain at city hall to ensure the confidentiality and security of citizens’ personal information. Furthermore, with the exception of classified information, Taipei City Government can provide related feedbacks to MediaTek Research as reference for future fine-tuning and improvements, utilizing actual use experience to promote future development of the language model.


(About MediaTek)

MediaTek Inc. (TWSE: 2454) is a global fabless semiconductor company. It is a leader in markets spanning smart mobile device, smart home application, wireless connection technology, and IoT product. There are roughly 2 billion devices and end products equipped with chips made by the company entering markets around the world every year. MediaTek pursues the innovation of technology, hoping to provide highly-efficient, low-power computing technology and advanced multimedia functions for smartphones, tablets, smart TVs, STBs, wearable devices, and automotive electronics. MediaTek inspires to make tech products more widely available, because the company believes that technology can improve the life of mankind and connect them to the world, while all people have the potential to access infinite possibilities through technology. For more information, please visit the company’s website: https://www.mediatek.tw 

(About NTUT)

The collaboration project is coordinated by director Lee Da-sheng of NTUT’s AI Digital Transformation Research Center (DTRC). The Center currently comprises 7 full-time assistants and 1 post-doctorate researcher, leading a team of 5 doctoral students and over 20 graduate students to conduct research work in the fields of AI and digital transformation in energy. Over the past 5 years, DTRC has published 24 papers across journals cited in the Science Citation Index, with research topics ranging from chiller plants temperature control to universal workflow of AI for energy saving. The Center’s research is highly-praised by the industry. It has also undertaken a total of 35 industry-academia collaboration projects over the past 5 years, receiving project fundings reaching NT$80 million. Over the past 5 years, the Center has carried out technology transfers to industries in areas such as machine failure detection device and method, helping it secure an amount exceeding NT$3.5 million.  In addition to being one of the research institutions receiving a high mark in the assessment conducted by NTUH, the Center performs on par with other digital transformation research institutions in both Taiwan and abroad. For more information, please visit the Center’s website: https://www.ntutdtrc.org/