For the manufacturing industry, replacing manual visual inspection with automated optical inspection is common, especially when the yield of 3C or semiconductor products is high General automated optical inspections often face the bottlenecks of insufficient defect samples and difficulties in qualitative and quantitative recognition Using AI deep learning for image defect detection has become increasingly significant AI detects minute defects, ASUS makes smart manufacturing 'visible' 'Initially, we hoped to promote upgrading with our 3C supply chain partners steadily, assisting the industry to enhance and face international competition,' said Chang Quande, ASUS Global Vice President and Co-General Manager of the Smart IoT Business Group ASUS Smart Solutions Business Unit uses AI deep learning to perform various workpiece defect detections, and layout after accumulating experiences is a priority task 華碩全球副總裁暨智慧物聯網事業群共同總經理張權德 For metal component manufacturers, detecting defects on surfaces is relatively difficult due to the reflection of light, which often causes actual defects to be overlooked Mastery of optical properties and the specifics of component surfaces is crucial The ASUS Smart Solutions Business Unit not only has AI experts but also a digital imaging technology team with unique post-processing skills and strong augmentation capabilities They can achieve correct defect data collection and train AI models efficiently even with a very small number of defect samples 'General optical inspection accuracy is about 85-90, and high precision-seeking manufacturers would not use it, as it implies a -10 defect misjudgment,' said Chang Quande Whereas manual visual inspection has an accuracy rate of about 93, it is labor-intensive and carries occupational hazard risks ASUS has now enabled AI to achieve 98 accuracy, fully capable of replacing manual inspections and certain traditional optical inspections Previously, it took three people to manage quality control across three production lines, now only one is needed In recent years, many manufacturing industries have been returning to invest in Taiwan Major metal structure stamping plants have also committed to establishing new factories ASUS has designed their three-in-one defect detection stations, capturing images through edge computing, uniformly training an AI model, and utilizing the same AI inference workstation to perform defect detection calculations Quality control stations across various production lines now monitor these processes in real-time Previously, three production lines required three people for quality control now, only one is sufficient, increasing the detection rate from 93 to 98 and reducing costs by 5 Accompanied by the reallocation of human resources, the stamping plant has achieved smart manufacturing and has broken the curse of increased production costs due to returning investments ASUS IoT 應用產業 除了金屬機構件之外，塑膠成型件、印刷電路板等電腦周邊元件生產業及系統組裝業都能運用AI 深度學習影像瑕疵檢測做高精度品管，目前也有半導體業正在優化導入華碩AI 深度學習影像瑕疵檢測，以補足自動光學檢測在晶圓層所抓不到的瑕疵，盼藉由AI的助力突破良率瓶頸，降低人工目測或自動光學檢測已知的誤判所造成的損失，更能利用人工智慧大數據針對品質瑕疵種類做統計分類以歸納出瑕疵形成原因，從源頭改善進而減少製程瑕疵。「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

How should one inspect diamonds that are only half the size of human hair The answer is to use AI for the inspection Zhihe Technology has developed the world's smallest diamond sorting machine By integrating AI and machine learning, along with accumulating large sample data, the system becomes smarter, increasing the yield rate from 70 to 96 within two years Moreover, by incorporating AI technology into the laser cross-cutting machine capable of handling 500-nanometer laser machining optical measurements, it becomes the first laser machine globally implanted with an AI system 2016年從中國大陸回流的智合科技，是一家專門提供微測距影像量測系統服務的新創公司，其主要的服務項目包括微米級m甚至奈米nm級的微測距應用、半導體檢測、鑽石相關產品、高米密度刀具量測及非標準件檢測。 Zhongxuan Li, General Manager of Zhihe Technology, has over a decade of experience in AOI inspection After returning to Taiwan from mainland China, he established Zhihe Technology At that time, semiconductor processes evolved from 8 nanometers to 3 nanometers Due to high difficulty levels in processing, Li saw a significant market opportunity Plus, AI's development accelerated after Google released the TensorFlow software in February 2017 TensorFlow is an open-source software library for machine learning applications in various perception and language understanding tasks World's Smallest Diamond Sorting Machine - Improved Yield to Over 96 Consequently, combining his expertise in AOI and AI, Li chose the diamond sorting machine as their first proving ground Zhihe Technology assisted a major industry company specializing in grinding, cutting tools, optics, wafer refurbishment, and other precision industries with automating human visual inspection process of diamond operations This company's star product, the 'Diamond Disk,' uses diamonds the size of half a human hair Previously, the company used traditional visual inspection, employing over 80 workers per production line, many of whom were foreign labor or older employees The process was time-consuming, output was low, and labor costs were high Most importantly, their semiconductor clients demanded automation, digitization, and increased precision for diamond inspections With Zhihe Technology's help, they supported a semiconductor equipment supplier in fully automating their manual diamond inspection process They helped create the world's smallest AI-equipped diamond sorting machine, increasing the yield rate from below 70 to over 96, gathering millions of diamond data points per day 靠AI技術打造全球最小的鑽石篩選機 運用AOI及AI交互應用，大幅提升鑽石篩檢效率 此外，智合科技在2019年10月與雷射設備廠商雷科科技共同合作開發雷射十字加工機，簡單說，就是在只有髮絲二分之一的尖點上，打上十字，其精度及準度的要求更高。智合採用AOI方式自動標注，測量位置與角度對位及加工高度，再運用AI訓練位置與角度估算核心，反覆校調後，耗費4個月時間，開發出全球第一台植入AI系統的雷射十字加工機，有了AI技術的加持，讓原有的雷射機價格翻了三倍之多。 智合與雷科科技合作，共同打造全球第一台AI 雷射機 Three brilliant methods to make Auto-AI digital transformation so easy Zhongxuan Li shared that Zhihe Technology is able to quickly integrate AI technology and develop Auto-AI, allowing enterprises to rapidly adopt and smoothly transition into digital transformation There are three main methods of implementation Method 1, Simplifying training issues with an automated labeling platform Use cameras to collect data from machine manufacturers, replace manual labeling with automated labeling, and progressively train to improve accuracy The simpler the problem, the less data is needed for training Method 2, Parallel advancement of AOI and AI In smart manufacturing processes, relying solely on either AOI or AI cannot achieve everything First, AOI should be used to mark features and distinguish between good and defective parts, followed by AI for labeling and training Using them in tandem enhances their effectiveness, and as the training data accumulates, the proportion of AOI decreases while that of AI gradually increases Method 3, Enhancing the integration capabilities of embedded system peripherals Establishing new computation platforms embedded systems or IPC platforms continuously enhances the computing power of AI, thus lowering the industrial threshold for AI applications 為降低AI使用門檻與成本，智合科技建立自主開發核心-Auto-AI又稱為傻瓜系統，目前已經跟國內知名工控電腦大廠進行合作，提供使用者更簡易的AI 使用環境。李忠軒表示，台灣是全球最適合作AI系統的國家，擁有超強的電腦設計能力與系統整合能力，若能再加上軟核心平台，將可大幅提升AI落地應用的實證。 AOI與AI交互並行，將AI應用落地時程大幅縮減 智合科技有研發能力相當強的機械控制及AI演算法的專業團隊，主要是公司的薪酬制度不同，智合將70的利潤分享給員工，讓員工共同享受公司的成長果實，因此能吸引優秀人才投入，即時協助解決客戶痛點，在不更新設備的情況下，藉由AI技術的導入，提升原有設備價值。李忠軒也自許，智合將從純粹業務推銷性質的設備商，轉變成為工業升級服務方案商，並將客戶的滿意度與安全感，轉變成為市場行銷上的卓越口碑。 智合團隊，圖左二為總經理李忠軒「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

What should small and medium-sized businesses or startups that want to export their products do when they lack PR resources, media exposure, and journalist contacts SparkAmplify, a company that builds global market platforms using big data, has created a precise media marketing platform SaaS that aggregates data from over 80,000 global media journalists With AI technology, it analyzes data and generates press releases within 20 minutes, matching them with accurately targeted international journalists to greatly increase exposure and achieve marketing goals internationally SparkAmplify's main service is a brand-media matching marketing SaaS platform Since its launch in 2018, it has continually analyzed international media trends and has already analyzed over 3 million international media reports, helping more than 1,200 companies from 25 countries achieve precise media exposure It has partnerships with major events such as CES and Computex, as well as famous incubation accelerators like TechStars, BootUp, Taiwan's TSS, Garage "Media are searching for news, companies are searching for media" By applying AI data, a balance has been found Jian-Qun Li, founder of SparkAmplify, explains, "From observing the demands of both suppliers and consumers in the media marketing market, there's a rigid demand for a platform that matches 'brands with journalists' based on both parties’ needs" Thus, SparkAmplify utilizes machine learning Logistic Regression algorithms to filter specific categories of news text and uses the LDA topic discovery algorithm to identify the hottest news trends, rolling out the 'AI Exploration of Media Trends' service Generate AI Press Releases in 20 Minutes to Find Suitable Media This system service only requires three major steps to disseminate the products or services of brands, small and medium-sized enterprises, and startups on the international market through international media coverage Step One, Material Preparation SparkAmplify sets up a dedicated brand page where brand managers prepare and upload complete materials including company profile, product names, service features, images, related product diagrams, etc步驟二、品牌故事撰寫：透過專家系統及運用機器學習Logistic Regression邏輯回歸演算法，將特定類別的新聞文本篩選出來，並透過主題探勘演算法LDA，找出最熱門新聞趨勢，系統會自動按結構、格式、片詞、文法、關鍵字等等，在短短20分鐘內自動生成AI新聞稿，再加以人工優化。步驟三、精準推薦：將公司及產品介紹、新聞稿等，媒合國際媒體共8萬名記者，將對的主題推薦到對的記者身上，主動提供記者報導素材，以增加媒體露出及曝光機率。 AI探勘媒體趨勢服務協助品牌公司精準國際曝光 Jian-Qun Li points out that traditional methods of gaining media exposure include holding press conferences or distributing press releases widely However, at international exhibitions, brand owners and small and medium-sized business leaders might not have sufficient PR resources Additionally, understanding industry trends and journalists' reporting preferences poses a significant challenge Aside from the challenges of data collection, extracting meaningful insights and trends can often be ineffective, time-consuming, and labor-intensive The 'AI Media Trend Exploration' technology can effectively and accurately collect data, use text mining and machine learning to unearth underlying information, and, by executing periodically, keep track of market changes to products 鎖定科技新聞領域 協助品牌業者精準曝光 善於資料分析的李健群，運用媒體大數據的分析技術，打造以機器學習進行分析的行銷系統平台，專攻歐美市場數據行銷決策與社群行銷，幫助行銷能力不足的的新創團隊，或有想要獲得國際媒體青睞的品牌業主，能以大數據分析找尋適合投放的媒體。 在AI技術的應用上，安普樂發使用NER命名實體識別技術，Named Entity Recognition技術來增加不同的屬性。例如人、組織、產品等，最後再透過知識圖譜Knowledge Graph建立屬性之間的關係，才能迅速達成預估目標。 由於新聞領域五花八門，包括財經、科技、政治、社會、運動、娛樂、美食、時尚設計等，資料數量眾多，但受限於儲存等資源，無法一一掌握，安普樂發將重點擺在科技新聞領域，與CES、Computex等大型國際科技展緊密結合，提供參展商在公關媒體上操作的資源，爭取國外媒體曝光機會，負責找對的媒體將品牌效益傳達、延伸出去。 三步驟完成媒體精準投放流程 SparkAmplify 商業模式主要為訂閱制，每月收取399美元，透過簡單步驟即可輕鬆完成品牌與媒體的對接服務。至於除了英語之外，未來是否會推出中文服務李健群表示，要跨到落地的語系需要重新建立一套模型，中文又比英文要複雜許多，處理過程要刪除非常多的雜訊。然而，因應中文化的需求日益殷切，未來在資源配置足夠的情況下，有機會也會推出中文服務。 SparkAmplify 團隊 SparkAmplify 創辦人李健群「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

Does AI cost a lot Is importing AI time-consuming and labor-intensive How to build consensus within the enterprise and build a solid AI team All the above problems are common problems faced by enterprises in AI digital transformation Huang Mingshi, the founder and CEO of the AI startup Kaiyuan Intelligent Manufacturing Company, provides customized AI solutions on a "subscription basis" so that traditional enterprises that want to undergo digital transformation can quickly introduce AI solutions The term "Lean Production" appeared for the first time in the book "The Machine That Changed the World" in 1990 When it comes to business operations, there is no waste of resources Phenomenon, the process operates smoothly and creates the most profit with the minimum investment AI subscription service solution to quickly assist in importing tools Huang Mingshi graduated from Jiaotong University and received a PhD in Electrical Engineering from Penn State University in the United States He also worked for a start-up company in Silicon Valley for five years He was the chief data scientist and led a team of 10 people to develop multiple AI projects, including real-time image recognition , 5GAI, prediction system for dynamic expansion of cloud resources, etc Huang Mingshi returned to Taiwan to start his own business and established Open Source Intelligent Manufacturing in May 2019 The development direction is to promote AI subscription services He hopes to make AI practical and practical, help small and medium-sized enterprises with AI application needs, and accelerate the cultivation of practical capabilities of AI talents Huang Mingshi believes that the scale of AI project software and hardware equipment, which can easily cost NT2 to 3 million, is an "unbearable burden" for small and medium-sized enterprises with limited funds and human resources However, for those who lack For small and medium-sized enterprises with limited resources, AI can indeed solve the problem of automation, reduce costs and improve efficiency for enterprises in a short period of time, and is an indispensable tool for digital transformation Therefore, Huang Mingshi follows the principle of "lean production" to prepare AI digital transformation tools for small and medium-sized enterprises, using small projects to get started, from corporate health clinics to identify problems, corporate training, consulting to providing AI model solutions Done within a year In the process of project promotion, we assist the company's middle- and high-level managers to empower them and help them understand the company's pain points, what problems AI can help solve, and introduce benefit analysis By then, it will be relatively easy to introduce medium and large-scale AI projects Open Source Intelligent Manufacturing’s customized solution for the legal industry is the development of a “food advertising text recognition and analysis” tool A medium-sized law firm wants to help clients solve the problem of "identifying food advertising violations" According to statistics, there are more than 4,000 cases of advertising violations in the food industry every month The traditional method is to assign 2-3 lawyers to search for exaggerations or claims of efficacy in food advertisements published by major media The cost of each lawyer is calculated at NT5,000-10,000, which means the overall cost is considerable However, by collecting relevant information through the crawler system, and then importing AI technologies such as natural algorithm NLP to develop food advertising text recognition and analysis tools, the recognition rate can reach 90, which can also greatly reduce personnel costs In addition, Kaiyuan Intelligent Manufacturing has successfully applied face recognition to applications such as smart tourism and smart door locks, achieving an accuracy of 95 It has also used graphic recognition to help digital advertisers achieve the function of AI image removal , reducing the time designers spend on repetitive memorization by more than 80 It is worth mentioning that for designers, it often took 2 hours to memorize photos in the past Using the AI model, 1,000 photos can be memorized in 10 seconds, which is amazingly efficient This means that designers do not need to spend too much time memorizing photos, and can use their time to come up with creative ideas When photos are needed, they can use AI technology to quickly memorize them for ordinary consumers, when making presentations When designing PPT, you can also use photos that have been reversed to speed up the presentation production time In the future, it will also be connected to Google Flickr's personal photo album or image search, so that you can directly memorize the required photos and complete the task in one go At this stage, the open source intelligent manufacturing project is cooperating with APP manufacturers to remove the photos and create a free website to benefit more people working in the design industry Open Source Intelligent Manufacturing develops an AI model for hair back removal, the effect is comparable to that of professional designers In the medical industry, Kaiyuan Intelligent Manufacturing has also developed obstetrics and gynecology organ image recognition technology and conducted education and training in schools to help students make correct judgments Kaiyuan Intelligent Manufacturing also cooperates with the Taiwan Suicide Prevention and Control Association to use AI models to find people who are emotionally distressed, have depression, or have negative emotions and comments on online forums such as PTT and social networking sites such as Facebook, and design Prepare a suicide risk assessment and submit relevant information to the Taiwan Society for Suicide Prevention and Control to prevent it before it happens and reduce possible tragedies Four-step import method to complete the goal within one year In order to help small and medium-sized enterprises achieve the goal of AI application through subscription services, Kaiyuan Intelligent Manufacturing hopes to use methods to find enterprise pain points in the shortest time and at the same time enhance the commercial value of AI applications The methods are as follows 1 AI Discovery Workshop Guide companies to explore their needs through workshops or corporate health clinics 2 Enterprise AI training The introduction of AI requires the full support of the company's senior managers and the consensus of all employees to be successful Through a one-month training, it can quickly transform into an AI-empowered enterprise 3 AI consulting What is important is the technical feasibility assessment Not a set of AI solutions can solve any problem Different AI models must be established according to the different needs of the enterprise in order to prescribe the right medicine 4 Subscribe to AI solutions Four steps of import method Huang Mingshi pointed out that the services provided by Kaiyuan Intelligent Manufacturing are not a single solution, but customized AI solutions for enterprises There is no problem with talents such as AI algorithms What is more difficult is how to market this set of consulting services to customers in a simple way Fortunately, Kaiyuan Intelligent Manufacturing has participated in the "AI HUB" and "AI GO" projects of the Industrial Bureau of the Ministry of Economic Affairs Through the method of "industry raising problems and talents solving problems", we can understand the needs of enterprises, solve problems accordingly, and launch customer services Customized AI solutions The open source intelligent manufacturing team won the excellence award in the AI GO problem-solving competition The picture first from the right is the founder Huang Mingshi 「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

Did you know a single golf ball can have up to 28 defect inspections Manually, one can inspect 500 balls in an hour, but with AI, up to 6,000 balls can be inspected in the same time Huiwen Technology has developed AOI Automated Optical Inspection technology that achieves a miss rate of 01, which is ten times better than human inspection Besides the golf ball industry, Huiwen Technology's AOI inspections are also being introduced to the textile industry and others Geng Cheng Lin, the founder and general manager of Huiwen Technology, has been an expert in artificial intelligence AI since 2013, recognizing the future potential and explosive power of deep learning DL and AI-based image recognition AOI has always been a strong demand in the manufacturing industry, mainly to improve product quality for business owners, stabilize the quality of delivered goods, and use data from AOI inspections to improve processes, thus creating a virtuous cycle and further cost reductions Due to uncontrollable factors such as human eye fatigue and inconsistent standards, inspection encounters bottlenecks The limit of human inspection miss rate after training is about 1-2, and the situation worsens over time AOI is a stable and capable of mass inspection device, achieving a miss rate of 01, which is ten times that of human eyes, implying a detection rate of 999 Of course, AOI also results in a 5-10 over-inspection rate, which can be further screened manually With the help of AOI, the burden of quality inspection is reduced, saving a considerable amount of labor time AOI Golf Ball Defect Inspection, Inspection Capacity Increased 12 Times per Hour The first litmus test of Huiwen Technology's AOI technology was on golf balls With their highly reflective, uneven surfaces, golf balls were previously inspected manually for defects A tiny golf ball can have up to 28 defects, and traditionally, only 500 balls could be inspected per hour A major domestic golf ball manufacturer, meeting the demands of Japanese customers, introduced AOI inspection two years ago The high-speed, high-precision AOI system combined with AI deep learning image recognition technology conducts defect detection on golf ball surfaces, fully automates the feeding and outfeed process, replacing manual recognition of missed defects, and can immediately record defect conditions and report back, inspecting up to 200,000 packs of golf balls per year per machine, greatly enhancing customer satisfaction However, this step took Huiwen Technology more than two years Golf Ball AOI Recognition Image Golf Ball AOI Recognition, 28 Surface Defects Unveiled Lin Geng Cheng says, from data assessment and consulting, followed by data organization and tagging, selecting and verifying AI algorithms to AI training services, the golf ball data is like starting from zero, accumulating one by one Thankfully, with full support from golf ball manufacturers, the efforts have finally bore fruit With AI inspection, while manually one might inspect 500 balls in an hour, AI can handle 6,000, achieving effectiveness 12 times greater Unlike other companies, Lin believes that AI needs to delve deep into domains to scrape professional data since only with such domain data can AI perform well Therefore, the company starts from individual projects, rather than setting an AI product from the beginning Without quality data or a focused domain, the best algorithms cannot succeed in AI Over the years, Huiwen Technology has accumulated project experience, gradually developing products while focusing on domain data and providing the latest AI algorithms to customers, growing together, creating a tighter collaboration, which is why, different from external fundraising, Huiwen's investors are customers or partners Evaluation to Official Launch AI Introduction Requires Six Phases The projects undertaken by Huiwen Technology are divided into several phases 1 Evaluation period, 2 Initial Validation POC period, 3 Data Collection period, 4 Repeated Verification period, 5 AI Positive Cycle period, 6 Official Launch The evaluation period involves understanding and assessing the Domain conditions of the demand side beforehand, followed by POC verification, extensive data collection after POC, entering repeated verification stage, and finally allowing AI to enter a positive cycle phase, achieving a certain level of effectiveness before the official launch Generally, a project takes about six months to a year to develop However, with more familiar PCB AOI projects, the first two stages are skipped, starting from data collection, thus significantly reducing the time 'Regardless of this project or others, common questions from customers are 'How much data is enough When will AI learn' Facing such questions, Lin points out that the reasons for these questions are 1 The inexplicability of deep learning technology, as it is a black box 2 Generally, customers lack the concept of AI technology Thus, the company must patiently verify data repeatedly, identify the data needed by AI, accumulate and test it, and clarify and resolve all Domain conditions, which requires a lot of time and patience During the AI introduction process, customers have high expectations of integrating AI services, thinking that it can immediately replace human labor Lin points out that this is not the case the real value of AI lies in accumulating large volumes of high-quality data, which is then transformed and analyzed to establish AI training and verification models to fully address problems generated by manual processes Apart from inspecting golf balls, Huiwen Technology is currently targeting the textile industry for items such as fabric and shoelaces, and many industries have conducted POC trials through Huiwen, including the semiconductor industry, PCB industry, and other traditional industries AOI Fabric Defect Detection, Top Image Shows Before AOI Inspection, Bottom Image Shows After AOI Inspection Lin Geng Cheng indicates that the most difficult aspect of entrepreneurship is nurturing talent and customer recognition customers often demand quick results, not realizing that AI adoption requires data accumulation and repeated verification, processes that cannot show results in less than six months Affected by the COVID-19 pandemic, the trend of globalization and centralization of the manufacturing supply chain has been disrupted, replaced by 'short region' supply chains, suggesting small, beautiful factories will flourish everywhere, potentially bringing new opportunities for AOI Lin notes that high automation indeed offers opportunities for automatic inspection AI, however, relatively high capital investments, including automation equipment, mainframes, GPUs, and sufficient AI maintenance talents, are burdens that small and medium enterprises or small factories cannot bear, requiring government financial resources and input to facilitate smooth transformation「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

In the era of AI, whether it's smart manufacturing or smart retail industry applications, the most important first step is data collection, especially on factory machines where thousands of information streams exist It's crucial to define what information is useful and what constitutes useless digital junk from the onset If there is simple, lightweight, and low-cost software, it could help enterprises collect data from machines and analyze and predict to achieve traditional enterprises' digital transformation goals AI Commercialization Insufficient, Main Consideration for Enterprises is Cost-Benefit Located in Xindian, New Taipei City, the headquarters of Beier Electronics Asia Pacific is a technology company committed to industrial automation and high-standard data communication Its main products include Industrial IoT Gateways devices linking two different network systems, transmitting data to other networks with similar functionalities but different structures, War Room Visualization software, HMI, with core technologies being controller communication drivers and visualization software Lee Li-Wei, the deputy director of product management and support at Beier Electronics, who has worked at Advantech and Xinchang Company, and familiar with Industry 40 operations, stated that when assessing whether to introduce AI, the most important consideration is cost and benefit From his observations over the years, the level of AI commercialization is still very insufficient because AI is often customized, not mass-reproducible, hence naturally expensive Only if AI moves to vertical applications and standardizes products, like using AOI for defect inspection or predictive diagnostics for motors or tools, can such AI possibly be commercialized and cost-effective X2 pro提供各種高性能工業用人機介面 At this stage, digital transformation challenges faced by factories include 1 Data collection difficulties 2 Factory equipment needing updates, which consumes time 3 High costs For large enterprises, sufficient funds and human resources allow them to introduce AI and undergo digital transformation projects As for SMEs, limited by resources, the key factor is cost-benefit, determining whether to introduce AI Existing software that assists in data collection and provides decision-making visualization software may meet current practical needs Once business owners see tangible benefits, they can further consider the need and incentives for adopting AI As a professional manufacturer of human-machine interfaces in Europe and America, Beier Electronics provides visualization software that fetches controller data on production lines and machine tools, as well as IT data integration services Using AI technology optimized for productivity and quality management can solve data fetching and integration issues For example, the factory war room displays the day's factory data and even real-time financial reports on a large screen wall Factory decision-makers can then use the integrated information and war analysis converted by visualization software to make decisions about production, marketing, inventory management, and procurement preparations Three major advantages of the war room low cost, easy maintenance, and mass reproducibility Compared to a typical factory war room, Beier Electronics' war room service has three major advantages it offers a low-cost, packaged visualization software it doesn't depend on engineers for maintenance and the war room can be easily commercialized and mass produced, which also accommodates future expansions from automation to IoT devices Lee further explains that AI primarily retrieves data for assessment, unlike automation, which demands real-time reaction It can tolerate a slower data fetching speed, still within milliseconds Human-machine communication doesn't need a special interface, so it can be decoupled from existing controllers without program changes, directly interfacing with existing hardware on the production line to fetch data, using existing software for data reading and analysis, aiding in decision management, and carrying out factory digitization upgrades As for whether Beier Electronics will introduce AI algorithm technology to provide users not only with data collection but also analysis and prediction services Lee stated that Beier Electronics considers three aspects Firstly, data collection is absolutely crucial when introducing AI At the same time, it must be done without increasing costs or changing site equipment for high customer acceptance Secondly, what problems AI aims to solve must be clearly defined Beier Electronics' clientele includes PLC programmable logic controller vendors, including Delta Electronics, Yonghong, Alliance Automation, Shilin Electric, and international giants such as Siemens, Rock weld, and MITSUBISHI However, the ultimate customers are PLC users, covering industries beyond semiconductors, including petrochemicals, 3C manufacturing, automotive manufacturing, power generation, and risk control The domain knowledge covered is extensive Given limited resources, whether to extend services to AI is still under consideration However, if professional specialization can be implemented, Beier Electronics plans to arrange an industrial ecosystem, introducing strategic partners to assist customers toward AIoT goals Deputy Director Lee Li-Wei, Product Management and Support Department at Beier Electronics「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

A tiny chip capable of driving AI algorithm speeds by nearly a hundredfold, Fansi Data's team is dedicated to software and hardware integration, providing industries including finance, smart healthcare, and smart manufacturing with a cost-effective, high-efficiency way of introducing AI and rapidly undergoing digital transformation In recent years, artificial intelligence has been highly prominent however, practical applications have been limited by high costs The enhancement of 'computing power' is crucial for breaking through the bottlenecks in AI applications Fansi Data's customized chip design and solutions can increase processing performance and effectively reduce costs, making AI applications in finance, healthcare, and manufacturing easy and feasible The company's core service is the high-performance hardware acceleration platform FPGA Fansi Data was founded in October 2018 by a founding team from National Tsing Hua University, National Chiao Tung University, and National Taipei University The company currently has 11 employees, including the chip design director Liu Wenkai from the IC design company Huirong Technology, who leads a 5-person IC design team They spent over a year developing the high-performance hardware acceleration platform FPGA, which became the company's core service Fansi Data integrates software and hardware to develop a high-performance hardware acceleration platform FPGA 'To bring AI to practical implementation, the challenges are cost and real processing situations Purchasing a standard set of NIDIA GPUs is expensive If we can adjust the hardware through customization, producing a setup tailored for use, the costs can be significantly reduced' Said Liao Yanchin, General Manager of Fansi Data, who additionally pointed out that most AI startups currently only have software engineers and lack hardware engineers Fansi Data excels in data handling and softwarehardware integration, has an excellent team, and can efficiently solve data issues while developing softwarehardware solutions tailored to customer needs Financial markets are notoriously fickle, as evidenced by the recent COVID-19 pandemic, which triggered a global stock market crash and was reinforced by program trading, leading to the unprecedented implementation of four trading halts in US stock exchanges within a decade This has significantly raised investors' risk awareness Zheng Zongyi, co-founder of Fansi Data, experienced in financial trading, points out that in financial markets such as stocks, futures, and warrants, 'speed' is often the key to victory Typically, the traditional stock trading process involves financial trading data flowing from the network to the mainframe, processed through combination software, measured in milliseconds ms, 10-3 seconds, with an average transaction completed in 20 ms System transaction processing speed, however, is at the nanosecond level ns, 10-9 seconds, and through the high-performance hardware acceleration platform FPGA, each financial matching transaction can be completed in microseconds, a significant difference that can lead to billions in trading gains or losses, and is a major competitive edge for proprietary traders Financial services in the domain of securities firms' proprietary sections, new types of financial product trading departments, and high-frequency traders or major retail traders In the securities market, market volatility is the result of a tremendous amount of data If the system operates at nanosecond speed, allowing you to see transaction information instantaneously, ahead by 01 seconds, you can make trading decisions before others even see the market data Service areas focus on financial technology and smart manufacturing The risk control systems of bank credit cards can also utilize AI integration acceleration, similar to regulatory technology domains Establishing an AI model can effectively identify risky credit card transactions and provide responses in a very short time, enhancing the security and smoothness of online transactions In the AI credit card risk control system, AI acceleration is also used through software integration Transactions are prevalent, and fraud is common, similar to regulatory technology domains By establishing an AI model, risky credit card transactions can be effectively identified, and responses given in a very short time, enhancing the security and smoothness of online transactions This includes financial transactions and credit card risk identification, all through chip-based transaction data analysis and risk management system direct acceleration calculations Financial transaction information acceleration solution Currently, many financial companies have their own IT departments, including data scientists, big data analysts, and AI algorithm engineers What is Fansi Data's advantage in the financial sector Zheng Zongyi points out that the IT departments in the financial industry are more 'users' of IT, not 'developers' of IT Moreover, understanding IC design involves high costs, and the financial industry does not need to maintain their IC design team The specialization is very clear, as Fansi simply develops models for the financial industry to adopt Considering personal privacy and data security, financial data is sensitive and often not easily accessible Fansi Data, by joining the financial technology innovation park FinTechSpace and with the assistance of the Institute for Information Industry, applies for the real-time transaction data and corporate annual financial statements, historical trading data provided by the digital sandbox, using it to group data, analyze, model, backtest, and propose AI risk warnings and other solutions for abnormal transactions and risk management Besides financial technology, Fansi Data also focuses on AI applications in smart manufacturing, such as developing smart image meter reading through image recognition methods, which can help businesses reduce equipment replacement costs and achieve higher accuracy In the process of customized chip design, data analysis, and softwarehardware integration, Fansi Data encounters difficulties in data and talent acquisition At this stage, through interfacing with the digital sandbox and utilizing resources provided by the financial technology innovation park, AI models are built regarding talent, a lean core team is established, continuously accumulating experience and building a robust entrepreneurial culture to face the ever-growing market demands From left to right Co-founder Zheng Zongyi, General Manager Liao Yanchin, and Chip Design Director Liu Wenkai「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

The introduction of the 'AI Medical Imaging Identification System' assists radiologists to conveniently and quickly complete image identification tasks, reducing their workloadDifferent Non-Invasive OptionsMedical imaging recognition is an important task for radiologists, who must make professional judgments based on patient examination data When a tumor is discovered, it is necessary to determine whether it is cancerous The possible methods include non-invasive medical imaging and invasive biopsy Although the invasive biopsy has a high accuracy rate, it also causes significant physical and psychological stress to the patientCurrently, imaging recognition can only determine the presence of tumors, not yet able to detect the difference between benign and malignant tumors To distinguish benign and malignant breast tumors, Datong World Science Company has assisted the Imaging Department of Changhua Christian Hospital, the first hospital in Taiwan to introduce the 'AI Medical Imaging Identification System' This system has increased the accuracy rate of artificial intelligence mammography in distinguishing benign from malignant tumors to 85, allowing for a shift from the original binary approach to a probability expression of BI-RADS gradingAI Medical Imaging Identification System Enhances Breast Cancer Diagnosis Accuracy to 85The AI medical imaging recognition system can assist radiologists in making quick readings Initially, it will target mammography When a tumor is detected, determining whether it is cancerous requires a pathological biopsy or mammography Pathological biopsy is invasive and, although more accurate, carries higher tangible and intangible costsMoreover, it helps improve the efficiency and accuracy of mammography readings Furthermore, optimizing the mammography reading process will reduce the workload on radiologists and decrease the waiting time for patients for examination results Additionally, with the aid of artificial intelligence, it helps reduce differences in radiologists' subjective judgments and prevent human errors, helping the institution to establish common standards and enhance collaborative efficiency among doctors from different specializationsCNN Convolutional Neural Network ModelIn addition to assisting doctors in making quick readings, here are summarized benefits of introducing the AI Medical Imaging Identification System1 Provides AI-assisted BI-RADS grading for mammography, helping radiologists in interpretation2 Optimizes medical imaging recognition processes, enhancing the degree of automation of existing procedures3 Uses local medical images to retrain models4 Adopts superior CNN models to improve accuracy and stability of the system5 Defines the relationship between BI-RADS grading and AI's readings of benign and malignant tumors transitioning from a basic dichotomy to a probability representation in BI-RADS gradingThe prerequisite for deploying artificial intelligence in medical assistant decision-making is that the accuracy must exceed 85, providing a valuable reference for radiologists With the support of artificial intelligence, the time for radiologists to interpret a single x-ray mammography image and assign a BI-RADS grade has been reduced to 50 of the original time, from about 10 minutes to under 5 minutes, offering an efficient and accurate AI-assisted outcomeChairman Baiyan Shen of Datong World Technology Co, Ltd「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」

The COVID-19 pandemic has spurred the popularity of delivery platforms such as Uber Eats and Foodpanda, creating an urgent need for smart dispatch systems Imagine if drivers could determine from their mobile phones or online platforms where there are no traffic jams, which roads have the fewest traffic lights AI could help plan the most suitable schedules, significantly improving logistics efficiency and reducing overwork With the flourishing of commercial activities, the logistics sector, which provides personnelgoods movement services, lacks smarter scheduling According to research by the international research organization Gartner, 97 of the global logistics industry does not use optimized software for effective planning Smart Scheduling Resolves Stakeholder Pain Points Let's first understand where the pain points lie among stakeholders in the logistics industry chain Employer's perspective In response to various types of delivery services, especially new types like food delivery, how to increase performance without expanding the fleet size Dispatcher's perspective Vehicle scheduling is very challenging, and the bosses demand increased efficiency, which is difficult to achieve without computers Driver's perspective Poor scheduling by the dispatcher leads to incomplete deliveries or traffic jams, often requiring overtime, or even accidentally running red lights, resulting in fines Addressing these issues, Zong Lan-Ken, founder and CEO of Singularity Infinite, states, 'All these problems are classical mathematical problems' Singularity Infinite's AIR Smart Dispatch Cloud Service is a cloud-based software service that resolves last-mile delivery scheduling and routing issues It addresses daily challenges faced by operators in managing goods, vehicles, and routes, enabling them to handle more orders with fewer vehicles Smart Scheduling System Schedule Zong Lan-Ken, who specializes in data science solutions for public needs and formerly served as an associate research professor at the Geographic Information Systems Research Center of Feng Chia University, founded Singularity Infinite in 2015 He aims to solve smart mobility issues using mathematics, statistics, and software technology The company's developed AIRouting optimization technique provides real-time traffic data and dynamic planning to assist operators in more efficient dispatching Singularity Infinite integrates real-time traffic and signal information and can handle high-frequency unconventional logistics models, such as gourmet delivery and electric scooter battery swap strategies For example, electric scooters must replace their batteries after every 50 kilometers If a scooter runs out of battery, the rider leaves it on the roadside The scooter operator must locate the depleted scooter and replace its battery To maintain effective operations, operators must keep the utilization rate of scooters between 80-90 For instance, in the Greater Taipei area with 10,000 scooters, maintaining more than 8,000 scooters on the roads at any time is crucial, yet without a smart scheduling system, high utilization rates cannot be maintained Following the system's introduction by WeMo in 2019, the utilization rate significantly improved by approximately 75 Effectiveness of AIR Smart Dispatch Cloud Service Introduction AIR Smart Dispatch Cloud Service has effectively increased the utilization rate by 75 Additionally, in the food ingredient delivery logistics, there have been notable results Traditional ingredient delivery companies need up to 25 trucks per day to transport fresh ingredients from produce markets, agricultural marketing companies, or seafood markets to restaurants After introducing the AIR Smart Dispatch Cloud Service, the number of trucks required per day was reduced to a maximum of 12, significantly cutting over half of the truck costs Singularity Infinite's team includes experts in mathematics, transportation, and AI technologies The traffic information used is from OpenStreetMap, supplemented with province-wide real-time traffic flow data to analyze congestion during different periods Additionally, future plans include using signal timing data to calculate which road segments have the fewest red lights and shortest red durations, to plan optimal routes, reducing the burden on logistics operators and drivers Singularity Infinite's team, the picture third from right is Zong Lan-Ken, founder and CEO of Singularity Infinite Besides logistics and transport, AIR Smart Dispatch Cloud Service can also be applied in container yard stacking, factory machine scheduling, project management, hospital bed allocation or operating room scheduling, and flight gate assignments among other areas Singularity Infinite employs two business models One involves customizing exclusive scheduling systems for clients, paid monthlyyearly on a pay-per-use basis the other involves system integration followed by revenue sharing with the client Fundamentally, Singularity Infinite provides APIs for integration, allowing operators to develop their own apps or provide services through websites In the entrepreneurial process, what are the most challenging aspects Zong Lan-Ken believes that entrepreneurship is a continuous series of multiple-choice questions, simplifying numerous questions into fewer choices, further simplifying each option to choose the correct answer Previously, it was mistakenly believed that 'technology can solve problems', but it was discovered that efficiency issues can not be solely resolved through mathematics, as the world does not operate this way In this ecosystem, 'who' will stop adoption due to 'whose opinion' For example, in the logistics industry, the most critical aspect of transporting goods is the driver, who needs rest If the system is introduced, and scheduling becomes completely transparent, drivers do not get time to rest The wrong introduction makes the system a tool for exploitation Hence, it is essential to consider human aspects, integrating rest times into the mathematical model to gain driver support Also, by knowing beforehand that a driver's home is near a train station, scheduling the last stop near the station lets the driver return home right after delivery These examples can significantly increase driver acceptance and greatly enhance the success rate of project adoption Zong Lan-Ken finally points out that data collection is crucial to the success of traditional industries' digital transformation in the future Without data, there is no data science, and no AI Singularity Infinite holds patents for automated data collection and recording, which can reduce data collection costs At the same time, the collected and stored data's high usability will serve as an important foundation for future intelligent logistics「Translated content is generated by ChatGPT and is for reference only Translation date：2024-05-19」