All Sessions：

Since 2018, the research team from NYCU has been actively advancing badminton sports science under the Ministry of Science and Technology’s Precision Sports Science Program. Over the past seven years, the team’s research trajectory has evolved from tactical data collection and analysis of elite athletes to the development of generative AI technologies designed for broader social engagement in sports. This presentation introduces CoachAI, a pioneering system that enables shot-by-shot microscopic data labeling, facilitating fine-grained performance analysis and intelligent feedback. It also highlights the team’s recent progress in smart badminton technologies and their efforts to extend these innovations beyond professional sports. Central to this initiative is to develop a photographic 3D sensing platform for smart sports spaces, aimed at making data analysis accessible in everyday athletic environments. The overarching goal is to embed sports data science into daily life, transforming how individuals interact with and benefit from physical activity.

Tactical and skill training play a crucial role in athletic development. With the support of artificial intelligence (AI) technology, it is now possible to track the ball and players to detect fine-grained events, helping coaches collect detailed statistics and infer each team’s tactics. Additionally, virtual reality (VR) technology can be leveraged to enhance both the effectiveness and experience of tactical and skill-based training. This talk will introduce modern systems that utilize AI and VR to help athletes conveniently gather valuable sports data and improve a wide range of skills.

This talk will share some practical AI landing cases with AI vision, AI audio, and Generative AI Technologies, such as AI smart sustainable agriculture, smart manufacturing, and smart health care. This talk will also introduce“AI maker”-- a platform for cultivating practical AI Talents by implementing AI vision recognition models and AI audio recognition models. The training program includes data (audio/image) collection, uploading, data augmentation, labeling, AI model training, download and execute AI models on end devices, such AR smart glasses, smart cell phone, AI tap stick, edge AI devices, robotics. The platform also provides tools to evaluate the performance of the trained AI models. We also build an alliance of sharing distributed GPU computing resources via Internet. The AImaker is a no-code platform and users can build and test their own AI models easily to accumulate practical AI experiences. 本專題演講將介紹 AI 視覺、AI 聽覺、生成式 AI 在實際產業落地的應用案例（包含智慧永續農業、智慧製造、智慧照護、智慧餐飲等等）以及 AImaker 實作型 AI 人才培育平台, 此平台專注於提供實作 AI 視覺模型、AI 聽覺模型、與 AI 生成式模型的服務。包含數據收集、數據上傳、數據增量（含 AI 生成數據）、數據標注（含 AI 自動標注）、AI 模型訓練、模型下載到多種終端設備（智慧型手機、AR 智慧眼鏡、AI 敲擊棒、edge AI 設備、機器手臂等等）進行模型的驗證與評量。本平台也透過網路整合多個夥伴單位的 GPU servers 形成算力共享聯盟，共同提供 AI 模型建置的算力需求服務。 此平台為 no-code 設計，不需撰寫程式即可輕鬆使用，創造並驗證（優化）自己的 AI 模型，累積 AI 產業應用的實務經驗。

This speech will illustrate E.Sun Bank’s efforts in combating fraud. We will explain the cash flow patterns corresponding to fraud and explain how to use AI to analyze cash flows. The features used include cash flow features and graph features. We hope to use these features to stably find the accounts used by fraud groups and set controls in a timely manner to block fraudulent cash flows. 本篇演講將說明玉山銀行在打擊詐騙方面的努力，我們將說明詐騙所對應的金流態樣，並說明如何使用 AI 來分析金流，其中所用到的特徵有金流特徵及圖形特徵，期望能夠以這些特徵來穩定地找到詐騙集團所用的帳戶，並能及時設控，以阻斷詐騙金流。

AI Agents and Digital Twins: Driving the Next Level of Autonomy in Manufacturing

Driving Smart Automation with AI, Digital Twin, and Physical AI, We’re building next-generation automation by integrating AI Robots、Digital Twin Infrastructure、Physical AI Applications、AI Nose SLM – a Smell Language Model. These technologies are unified under an AI-powered intelligent platform, enabling a closed-loop logistics and material handling system (R2S & S2R) for the semiconductor industry and smart logistics. From day one, the system is designed to make the right decisions early—delivering the optimal solution through AI-driven precision and feedback. 以 「AI、數位孿生、Physical AI」 為主軸，導入 AI Robot、數位孿生 AI 基礎建設、Physical AI 、AI Nose SLM 嗅覺語言模型等應用方案，並以 AI 驅動的智慧平台，提供智慧物流及半導體自動化物料搬運系統 R2S S2R 的 Close Loop 架構，一開始就做對決策，找出最佳解。

Recent developments in artificial intelligence and machine learning have remarkably advanced machines’ ability to understand images and videos, comprehend natural languages and speech, and outperform human experts in complex games. However, building intelligent robots that can operate in unstructured environments, manipulate unknown objects, and acquire novel skills – to free humans from tedious or dangerous manual work – remains challenging. My research focuses on developing a robot learning framework that enables robots to acquire long-horizon and complex skills with hierarchical structures, such as furniture assembly and cooking. Specifically, I present an interpretable and generalizable program-guided robot learning framework, which represents desired behaviors as a program and acquires primitive skills for executing desired skills. This talk will discuss a series of projects toward building this framework.

After Mobile,what is next ? XR Devices Categories and mainstream reveal detail about Core Technologies of AR smart glasses How AI-Powered AR Glasses spatial computing can transform the way you work, connect, and experience the world! Why AI needs AR Glasses：Not PC and Mobile AI make AR better、fast and powerful,AI need AR even more。AR give AI own Human sensor to access the realworld,no AI without AR

This talk explores how to use open-source LLMs models to build AI Agents to enhance productivity and efficiency across various industries. As AI Agents become a major focus by 2025, a wide range of applications will be widely promoted, providing enterprises with opportunities for challenge and upgrade. Open-source models have three main advantages: on-premises deployment enhances security protection, ownership of weights increases system stability, and fine-tuning creates a better user experience. Moreover, the open-source Breeze 2 from the MediaTek Research not only serves as a Traditional Chinese model but also targets AI Agents; we will open source a lightweight Agentic Workflow to help Taiwanese industries easily deploy Breeze 2 as an industrial assistant. This presentation will illustrate how to use these open-source models to create AI Agents through examples, including real-world cases from Cola Tour and AIA. 這個講題探討如何利用開源的 LLMs 模型來建立 AI Agent，以在各行業中提升生產力和效率。隨著 2025 年 AI Agent 成為一大重點，各類應用將廣泛推展，企業因此迎來挑戰與升級的契機。開源模型具備三大優勢：地端部署提升資安防護、擁有權重增加系統穩定性、以及可微調創造更佳使用者體驗。此外，聯發創新基地開源的 Breeze 2 不僅是一個繁體中文模型，更是一個對標 AI Agent 的模型；我們將開源一個輕量級的 Agentic Workflow，以助力台灣產業輕鬆部署 Breeze 2 成為產業助手。本演講將通過實例，包括使用可樂旅遊和 AIA 的實際案例，說明如何利用這些開源模型打造 AI Agent。