The Central Weather Administration (CWA) is the government agency responsible for earthquake monitoring and reporting in Taiwan. In order to reduce fatalities and property losses caused by earthquakes, a high-density real-time seismic network and the advanced earthquake early warning (EEW) system have been implemented and in operation. Currently, the CWA EEW system can issue the warning message in about 10 seconds after a strong earthquake occurring in or very near Taiwan. Starting from 2021, in order to further enhance the benefit of EEW to the practical earthquake hazard mitigation, the CWA has begun to develop the customized urban EEW system. By means of introducing the on-site earthquake early warning methods, and developing the artificial intelligence models. The systems, particularly for the big urban shallow earthquake, aim to further shorten the time of EEW to 7 seconds. That means it can give warning to the areas about 25 km away from the epicenter, and more people could be alerted before the destructive shaking.

This project aims to establish the first "1 Stop Solution Center for 5G XR Education Content Production" in Taiwan. It involves setting up a pioneering 5G O-Ran XR studio to produce instructional videos using XR stereoscopic technology and post-production effects. The content will be streamed in real-time to VIVERSE, allowing schools to receive it through 5G Private Mobile Network for immersive VR teaching across various devices. This innovative solution not only streamlines content creation but also addresses the urban-rural education resource gap and alleviates the common issue of insufficient teaching staff in remote areas.

LanCore Intelligent Aquaculture Cloud System is a high-end product used in aquaculture fishery, with intelligent, high-efficiency and power-saving effects. Its main features are: (1) Integrate more advanced self-development software to achieve intelligent control, which can remotely control aerators, feeders, water pumps and other breeding equipment. (2) The intelligent scheduling control mode supports timing setting, numerical value setting, and dynamic operation adjustment. The power saving effect is quite significant, up to 60%. (3) Multi-scenario scheduling and dynamic operation adjustment, high-speed operation is used at night to increase the amount of dissolved oxygen. During the day, the aerator can be adjusted to run at low speed to activate water quality and create an excellent breeding environment. (4) The grouping control system is ideal for cross-field and large-scale fishery users. It is easy to build cloud control and realize energy saving.

Handling abnormal complaints requires extensive experience and expertise. However, the resignation or retirement of senior employees often leads to knowledge loss, impacting organizational efficiency and competitiveness. Through vector models, Morale AI transforms unstructured data (such as emails, PDFs, and Word documents) into structured data and integrates it into databases, enhancing data usability. Multi-modal models combine diverse data types, including text and images, to systematically record and preserve experience. By leveraging Retrieval-Augmented Generation (RAG) technology and similarity search, the model can accurately categorize issues, identify the root causes, and recommend relevant cases, significantly reducing manual processing time and error rates. This systematic approach not only enhances response efficiency and enables real-time analysis and resolution of customer issues but also reduces repetitive work for employees, improving overall productivity and customer satisfaction.

P-Waver Inc. is the global leader of AI on-site earthquake early warning system and structural safety monitoring system, providing earthquake early warning before earthquake, intelligent disaster mitigation measures, and structural safety monitoring after earthquake.

The system developed by this project provides a smart solution for station including digital fence system, platform pedestrian flow analysis, animal tracking expelled system, track water situation analysis system and platform digital information signage. The platform comprehensively analyses the image of around the station, train coordinates, passenger entry and exit, and water depth etc. Which can be used for follow up review and improvement of abnormal events such as track intrusion, crowd congestion and road flooding to achieve optimal train operation configuration and class spacing control. And provide abnormal event warning and troubleshooting and passenger intelligence information services. After the integration of multiple subsystem through the IOC platform, the system will automatically alert an alarm to the traffic control center and announce the train near the area to remind the driver to pay attention to the situation and enhance the safety of the train and the people or cars in the area when the detection system alarms.

MiTAC has leverage the RFID technology and integrated with related solutionsto modify the warehousing and production management system. MiTAC's RFID Smart Manufacturing Solution has been applied in various industry, such as IPC, LED, Semiconductor, and textile industry. MiTAC has provided complete RFID application solutions these years from the inventory to production management system. We sees RFID as a game changer. Besides the RFID technology, the integrated application system is also important. MiTAC also develope the application systems. For each different scenario. We’ve developed a different system to work out a more complete RFID solution. We've also used supervised learning AI modules to develop MiTAC’s MiAION SeeR, a failure detection solution. MiTAC's strong software development ability is the biggest advantage.

Currently, the prognosis of critically ill patients in the intensive care unit (ICU) is often assessed using the APACHE or SOFA scoring systems to predict mortality rates. However, these systems have been in use for over 20 years, and the data they are based on is outdated. Today, clinical data is more abundant and precise, and advancements in machine analysis capabilities have made traditional scoring systems insufficient for meeting the demands of accurate predictions and personalized medicine. To address this, our team has developed a survival prediction device for critically ill patients. Using an AI algorithm model, it utilizes clinical data such as physiological, laboratory, and examination results to predict the patient's survival rates for the next 30, 60, and 90 days, providing a more accurate reference for prognosis assessment and medical resource allocation.

Smart Traffic Control System: Traffic Congestion Improvement Project, AI Image Recognition and Management System

Our platform, TradeTioh (Taiwanese for “found it”), focuses on the local Taiwanese market. Through customized matchmaking services and technological innovation, we offer two-way matching features, real-time communication tools, and data-driven optimization algorithms to help buyers and suppliers quickly establish cooperative relationships. Looking ahead, we will continue to enhance our platform technologies and expand into international markets, aiming to bring long-term value to Taiwan’s manufacturing industry. On the TradeTioh platform, buyers can submit their procurement needs, and our core AI-driven technology will analyze the requirements and process them through our big data backend. The system then recommends the most suitable supplier candidates. Buyers can use the built-in messaging tool to directly contact the selected suppliers. The entire process takes only a few minutes—powered by TradeTioh’s core technology—making it far more efficient than traditional methods such as browsing websites, flipping through directories, or relying on referrals.

Our AI server features several advanced patented functionalities: • Ensure the Privacy of Local Data: Patented trusted computing protection mechanisms ensure data security and privacy. • Protect Local AI Models from Theft: Prevents locally deployed AI models from being stolen, safeguarding corporate intellectual property. • Specialized LLM Compression Technology: Efficiently processes large AI models, enhancing performance and reducing resource requirements. Enables the use of NVIDIA A-series GPUs in place of H-series GPUs, significantly reducing hardware costs. • Reduce Enterprise Deployment Costs and Energy Consumption: For example, when training a 13B model, GPU requirements can be reduced from 7 H100 GPUs to 2 A6000 GPUs, significantly lowering costs and energy consumption. Our solution not only ensures the security and privacy of corporate data and AI models but also reduces the costs and energy consumption associated with AI deployment, making it an ideal choice for enterprise users.

The application of digital twin technology in building smart cities is becoming increasingly clear, serving as a crucial engine for driving urban transformation. In the development of smart cities in recent years, citizens can view the growth and development of the entire city, including population growth, new construction, and other major activities, in a 3D format, gradually emphasizing the concept of a virtual city. Digital twin technology can simulate disaster scenarios, including weather events and traffic accidents, assisting disaster response personnel in rapidly understanding the situation and optimizing the efficiency of disaster response. In recent years, with the advancement of aerial photography, simulation technology, and computing resources, The National Science and Technology Center for Disaster Reduction has leveraged advanced technology to create a mountainous flood inundation model system. This system enhances disaster warning and impact analysis, providing a comprehensive view of vulnerable settlements. It utilizes dynamic models and virtual-real integration, presenting alert information and disaster impacts through visual elements, contributing to informed decision-making in disaster situations.