BIOSTAR, a leading manufacturer of edge computing solutions, motherboards, graphics cards, and storage devices, proudly concludes a highly successful COMPUTEX TAIPEI 2025 showcase. This year's campaign drew significant attention from global media, industry professionals, and tech fans, highlighting BIOSTAR's strong presence across edge AI platforms, industrial edge computing, and next-gen AI PC solutions. Credit: BIOSTARA major highlight of BIOSTAR's COMPUTEX 2025 campaign was the official debut of the EdgeComp series, a new generation of edge computing systems developed to meet the demands of modern industries. Drawing on over two decades of expertise in industrial motherboard manufacturing and deep experience on developing industrial computer systems, BIOSTAR's engineering team has built platforms that are purposefully designed for smart manufacturing, smart retail, and smart city applications.Credit: BIOSTARAmong the featured edge AI platforms, the EdgeComp B850MT-H1 demonstrated its capability to handle real-time AI recognition across more than 100 video channels simultaneously. The EdgeComp N97-MX3, equipped with the MemryX AI NPU, demonstrated real-time environmental safety monitoring applications, including construction site safety and wildfire detection scenarios. The EdgeComp AI-NONXS was demonstrated live performing three simultaneous edge AI tasks: multi-object vision recognition, body pose detection, and a locally-run SLM AI assistant. BIOSTAR showcased how seamlessly integrates leading-edge AI NPU solutions from NVIDIA, DEEPX, and MemryX to deliver reliable, high-performance edge AI computing.As AI continues to reshape the future of technology, BIOSTAR is strategically positioned to lead with its close partnerships with industry giants like Intel, AMD, and NVIDIA. This synergy allows BIOSTAR to fine-tune hardware and software integration at every level, ensuring optimized performance, stability, and adaptability across a wide range of edge AI workloads. From automation system integrators to digital signage specialists, HMI solution providers, and edge AI developers, BIOSTAR offers tailored solutions that meet the specific needs of each client, delivering not just products but full-stack platforms engineered for real-world applications.Among the notable EdgeComp systems unveiled at COMPUTEX, BIOSTAR introduced the ultra-compact EdgeComp MU-N150 and the robust 2U rackmount platform EdgeComp MW-B670, highlighting scalable edge computing solutions for space-constrained deployments and data-intensive industrial applications. From the EdgeComp MS-1335U and MS-X7433RE to the MS-J6412, BIOSTAR showcased a comprehensive range of EdgeComp embedded systems, tailored to meet the diverse demands of smart manufacturing, smart retail, and smart city applications.Credit: BIOSTARBIOSTAR also showcased the EdgeComp MS-N97 and EdgeComp MS-J6412 in a DIN-Rail demonstration setup—an industrial-friendly design tailored for factories and compact spaces, enabling wall-mounted installation via rail systems to optimize space utilization.Credit: BIOSTARIn addition, BIOSTAR showcased its reliable range of industrial motherboards, including the BIW88-AHS with ECC DRAM support and the versatile BIQ87-AHS which designed for industrial automation. BIOSTAR also introduced the first NVIDIA Jetson Orin edge AI developer kit, the AI-NONXS, compact SBC BITWL-N150, ERX93-AXP, and RASL-D5DE, purpose-built for EV charging solutions.On the consumer side, BIOSTAR's booth also drew significant attention with its showcase of custom-built PCs and cutting-edge components. Two standout modded systems wowed visitors, one built around the Z890 VALKYRIE motherboard in a high-performance showcase rig complete with a dedicated highlight video and another fully water-cooled ARGB build powered by the B850MT-E PRO motherboard and AMD Ryzen 7 9800X3D processor. Running DDR5 memory at 8000MT/s, the system impressed with both looks and performance.BIOSTAR's consumer motherboard lineup at COMPUTEX 2025 featured a strong mix of Intel and AMD platforms designed for next-gen computing. On the Intel side, the Z890A-SILVER leads with support for Intel Core Ultra processors, PCIe 5.0 M.2, dual HDMI, DisplayPort, 2.5GbE LAN, and WiFi 7 module support. The B860M-SILVER and B860MT-E PRO offer mainstream flexibility with similar high-end features, including DDR5 support up to 256GB and 2.5GbE LAN. For AMD users, BIOSTAR showcased the industry's first B840 chipset motherboard, the B840MT-E, built for compact systems. The B850MT-E PRO and B850MT2-E DJ deliver high-speed connectivity, PCIe Gen5 or Gen4 M.2, and broad support for the Ryzen 9000/8000/7000 series, ideal for gaming and multimedia applications.Complementing its motherboard and system showcases, BIOSTAR revealed a strong lineup of consumer hardware designed for modern performance builds. The AMD Radeon™ RX 7000 series graphics cards were a major highlight, featuring the RX7900XTX, RX7900XTX, RX7800XT, RX7700XT, and RX7600OC, which deliver exceptional graphics performance for gaming, creation, and compute-intensive tasks. For storage, the spotlight was on the M800 PCIe Gen4x4 M.2 SSD, built for high-speed data access and efficient thermal control in compact, high-performance systems. BIOSTAR also showcased its GAMING X DDR5 memory modules, designed for next-gen platforms and featuring sleek heatspreaders and dynamic RGB lighting, which combine speed, stability, and style in a single package.Credit: BIOSTARLooking ahead, BIOSTAR is committed to driving innovation in the edge AI space, building on the momentum of its successful EdgeComp series. Through close collaborations with leading technology partners, including Intel, AMD, NVIDIA, DEEPX, MemryX, and others, BIOSTAR continues to develop next-generation edge AI platforms that are smarter, faster, and more adaptable to real-world applications. As global demand for AI-powered edge solutions accelerates, BIOSTAR stands ready to support its clients with advanced, reliable technologies rooted in decades of engineering expertise as a trusted edge computing partner.Connect with usWebsite https://www.biostar.com.tw LinkedIn https://www.linkedin.com/company/biostar Facebook www.facebook.com/BiostarHQ X https://twitter.com/BIOSTAR_Global Instagram www.instagram.com/biostarofficial YouTube www.youtube.com/user/BiostarTaiwan Availability http://www.biostar.com.tw/app/en/wheretobuy/index.php BIOSTAR VIP Care https://store.biostar.com.tw/customerABOUT BIOSTARBIOSTAR is a professional brand dedicated to the production of motherboards, graphics cards, computer peripherals, industrial PC, and AI computing solutions. Since its establishment in 1986, BIOSTAR has become a major supplier in the IT industry. In order to pursue the best quality and aesthetic design, BIOSTAR has invested heavily in ID design, global marketing as well as R&D. With an unceasing focus on enriching lives, BIOSTAR consistently strives to pioneer a fresh future, bringing solutions into reality.

As climate change escalates and significant disasters arise more frequently, Taiwan's public safety concerns continue to grow. The gas explosion at the department store in Taichung in 2025 underscored the critical issue that the "golden rescue time" prior to the arrival of trained rescuers is frequently underutilized. During the second "2025 AI Wave: Taiwan Generative AI Applications Hackathon," the "Revenge Seekers" team identified a significant social issue and provided a unique AI-driven solution, ultimately securing victory in the "Resilient Future" category of Chunghwa Telecom."Revenge Seekers" wins Chunghwa Telecom's Resilient Future award with an AI-powered emergency assistant that helps save lives in disaster zones. Credit: CompanyIn response to the frequent occurrence of significant disasters in recent years, the "Revenge Seekers" team has developed an intelligent system that is designed to aid non-medical individuals during emergencies. The purpose of this system is to improve the disaster response capabilities of the entire population by leveraging technological advancements. The team conducted a comprehensive analysis of significant accidents, including the 2015 New Taipei water park dust explosion and the department store gas explosion in Taichung earlier this year, and identified three urgent issues at the disaster sites that required immediate attention.First, the general public lacks basic first aid knowledge, preventing them from effectively using the "golden rescue time" before medical personnel arrive. Secondly, the inability to record and classify casualty information on-site in real time leads to the inefficient deployment of following medical resources. Thirdly, the prompt distribution of casualties is challenging to understand, significantly affecting the overall decision-making and resource allocation of governmental agencies during rescue operations. These problems become even more severe in mass casualty events.The team developed "First Aid Master - Public Edition v2.0" to resolve the aforementioned pain points. This first aid assistance system integrates generative AI technology. It is capable of assisting individuals without a medical background in conducting preliminary triage of the injured and providing situational basic first aid guidance through an intuitive interface. The system will categorize the injured into four categories based on the current condition of the injured as entered by the user: red (immediate treatment), yellow (treatment can be postponed), green (minor injuries), and black (extremely low chance of survival). It will then provide corresponding first aid suggestions for each category in real time, such as stopping bleeding, maintaining airway patency, and performing CPR.Furthermore, the rescue efficiency and resource allocation accuracy can be significantly enhanced by the system's ability to autonomously generate injury records and zoning information for the reference of professional medical teams that arrive later."First Aid Master" is built on Amazon Web Services (AWS) cloud infrastructure and leverages multiple innovative technologies to enhance the stability and responsiveness of system operations. The team utilized AWS PartyRock, a no-code, generative AI application builder, in conjunction with large language models such as Amazon Nova Pro and Anthropic Claude 3.5 Sonnet v2 as the primary AI engine. To enhance the professionalism and precision of answers, the team independently structured the emergency documents and developed a compact retrieval-augmented generation (RAG) system.Amazon CloudFront delivers content with low latency through its content delivery network with secure HTTPS connections; Amazon Route 53 ensures reliable and efficient DNS routing, while Amazon Elastic Compute Cloud (Amazon EC2) provides scalable compute capacity for the website and backend services. The open-source search engine OpenSearch retrieves first aid knowledge base information, and AWS WAF helps protect web applications and APIs from attacks. The system incorporates an AED map feature to help individuals quickly locate the nearest automatic external defibrillator, significantly improving rescue response time—critical for patients experiencing cardiac arrest."Revenge Seekers" distinguished itself within the Chunghwa Telecom "Resilient Future" group. The judges determined that this technology has the potential to provide immediate assistance at disaster sites and to function as an emergency education instrument during peacetime to disseminate first aid knowledge to the public. The system has substantial practical utility and can help Taiwan establish a robust civilian emergency network. Moreover, the system's real-time summary of patient information provides public departments with valuable insights for decision-making, which are expected to enhance the allocation of medical resources.

With the advancement of digital technology and the onset of the post-pandemic era, the human desire for emotional bonds has increased. The idol industry is experiencing substantial transformational challenges as fans grow discontented with unilateral idol worship and seek more interactive experiences. At the second "2025 AI Wave: Taiwan Generative AI Applications Hackathon," the "Disheveled in the Wind" team created an innovative solution merging physical dolls with AI technology, establishing a new paradigm for the idol economy, and won the Gamania Group Vyin AI's "AI Idol" category.Disheveled in the Wind wins "AI Idol" category with Aidol—an interactive AI companion that redefines emotional connection in fandom. Credit: CompanyThe "Disheveled in the Wind" team members are from Taiwan's National Tsing Hua University and National Central University. Their enthusiasm for the idol sector, combined with their robust technical expertise, led them to leverage generative AI technology to develop a sophisticated idol product capable of genuinely engaging with fans, addressing the market's need for emotional connection, and creating new value within the idol economy.Through comprehensive user research, the team discovered that most current idol products are unidirectional, showing a substantial lack of interactivity and genuine connection. Despite their profound devotion to their idols, fans are unable to establish authentic connections due to limited interactive resources. The user study findings reveal that fans are intrigued by the emotional depth, personality, and authenticity of their idols' voices. They anticipate receiving unique responses tailored to specific occasions, such as anniversaries and farewells, rather than mechanical, standardized interactions.In response, "Disheveled in the Wind" created an intelligent doll named "Aidol" that integrates interactive capabilities, vocal communication, and emotional memory functions. This device is advertised as an "AI idol that can accompany you" and has five primary purposes.The five key features include: first, the "share daily life" function, allowing users to upload photos to the cloud and receive feedback from idol about their lives; second, "personalized conversation memory," archiving previous interactions for a more authentic idol experience; third, "exclusive event reminder," automatically notifying users of birthdays, idol activities, and other important events; fourth, "multiple personality switching," offering various personality modes including energetic, gentle, dark, and comical; and fifth, "manual switch memory and interaction intensity," enabling users to regulate interaction pace and protect privacy. Through these features, "Aidol" delivers an unparalleled companionship experience while addressing the limitations of traditional idol products' one-way interactions.The technical architecture of Aidol is comprised of two components: cloud AI services and hardware.The doll features an integrated camera, microphone, speaker, sensor, and Bluetooth module that enable interaction through a mobile application. For cloud services and AI technology capabilities, the solution uses Amazon Bedrock, a fully managed service from Amazon Web Services (AWS) that makes high-performing foundation models, along with speech-to-text, text-to-speech, and natural language understanding services. The team also established a database of idol-exclusive quotes and catchphrases to enhance the naturalness of the interaction.The system's core employs Raspberry Pi to process sensors, networking, voice processing, and other modules, while also connecting to Gamania Group's API to access idol-exclusive sound resources. The prototype clip provided by the team in the competition clearly demonstrated the actual operation process of several activities such as shooting images, recalling discussions, and pushing events, and was well appreciated by the judges.This innovative idol product solution received widespread praise from the judges of Gamania Group Vyin AI's category. They commended the product for its comprehensive business plan and technically feasible implementation, noting its potential to create a unique companionship experience for fans while generating various revenue streams, including hardware sales, app subscriptions, and downloadable voice and appearance packages. Compared to existing market offerings, Aidol focuses on the emotional economy of idol fandom, demonstrates clear competitive advantages, and presents significant potential to become a major technological innovation in the idol industry.

Generative AI (GenAI) is swiftly revolutionizing corporate operations, product development, business models, and the overall ecosystem. According to a survey report published by Taiwan's Market Intelligence & Consulting Institute (MIC), in 2024, 19% of Taiwan's five major industries utilized GenAI or engaged in related activities, with the finance and insurance sector representing 25% and the manufacturing sector following at 22%. Amid the proliferation of Generative AI for developing AI assistants, some firms have found that their substantial investments in these assistants did not yield the expected results, leading them to terminate their AI projects and thus diminishing their overall competitiveness.Alpha Three wins with a smarter AI assistant—boosting accuracy through improved data chunking for enterprise knowledge Q&A. Credit: CompanyThe primary cause of the poor performance of AI assistants, according to "Alpha Three," the winning team in the "2025 AI Wave: Taiwan Generative AI Applications Hackathon" from Walsin Lihwa's "Smart Manufacturing" group, is the excessively small data chunking during pre-processing. This can easily disrupt the original document paragraph context, resulting in the AI model's misunderstanding deviations and a response content that is not sufficiently accurate.The team recommended that the "amount of text in a single PDF page" be used as the unit of chunk in order to preserve the natural paragraph structure and comprehensive context, as well as to prevent semantic discontinuity. The review committee unanimously recognized this method for successfully achieving three substantial advantages: "optimizing user query experience," "reducing the risk of hallucinations," and "enhancing semantic coherence and search and answer accuracy."Pre-competition training proves valuable; effectively utilizing AI tools to realize creativity"Alpha Three" utilized a steel standard inquiry as a test case and posed the question, "Does ASTM A276 steel grade 316Ti comply with the EN 10088-3 standard?" The AI system retrieved comprehensive information covering the chemical composition and standard specifications of steel grades. The content of the response is highly focused and accurately reflects the primary data. The AI system demonstrated extraordinary reliability in the application of enterprise knowledge by achieving a perfect score (1.0 out of 1.0) across the three metrics of "search relevance," "answer solidity," and "answer relevance."To achieve these results, the team utilized Amazon Web Services (AWS) to develop a comprehensive enterprise knowledge question-answering framework.PDF, PNG, JPG, and other file formats are uploaded to Amazon S3, the cloud object storage service, during the initial phase. The second phase is providing quick query services with a comprehensive language model and the Flask API. In the third phase, the team leverages Amazon Bedrock, a fully managed service that makes high-performing foundation models, to connect extensive language models, thereby improving scalability and reaction speed. Amazon Elastic Compute Cloud (Amazon EC2) is employed in the fourth phase to expedite API processing, thereby guaranteeing system stability and efficiency.The extensive system design includes data uploading, management, retrieval, and response, allowing users to easily submit inquiries and obtain prompt professional responses, which became a crucial factor in their victory.The "Alpha Three" team, comprised of recent information engineering graduates from National Taiwan University in 2024, observed that, despite their degrees in information-related fields, they were completely unfamiliar with contemporary mainstream AI tools in the face of the rapid advancement of Generative AI technology. The project was successfully completed within 30 hours, and the award was secured, thanks to the professional training provided by the organizer, which included a series of enterprise data workshops and AWS Generative AI workshops, as well as Walsin Lihwa's explanation of the steel standards.

IntroductionWith the increasing use of mobile devices, malware targeting smartphones and tablets has become more prevalent. Banking Trojans, in particular, are designed to steal banking credentials and financial information from mobile users.The modern trend in the Integrated Circuits industry is System on a Chip (SoC) and Microcontrollers (MCU), which integrate different discrete solutions, including security functions in a single IC. In particular, the Secure Element / Hardware Security Module (HSM)/UICC can be integrated into the SoC. The main motivations for this integration are reduced system cost, enhanced performance, and added-value functionality.The integrated security function in the SoC needs to meet the same security level as the discreet part. To address the security of integrated solutions and provide the industry with a unified set of security requirements to be fulfilled and clear to evaluate and asses, PP-0117, Secure Sub-System in System-on-Chip (3S in SoC) Protection Profile was developed.BackgroundCybersecurity statistics indicate that there are 2,200 cyber-attacks per day, with a cyber-attack happening every 39 seconds on average. In the US, a data breach costs an average of $9.44M, and cybercrime is predicted to cost $8 trillion by 2023.ENISA[1], in its "ENISA Threats Landscape 2022 Report", presented in several aspects that the segments which were affected the most were the Public Administration and the Finance sectors:Figure 1 ENISA: Reputational impact by sector[1] ENISA - European Union Agency for Cybersecurity, https://www.enisa.europa.eu/This figure points to the potential for negative publicity or an adverse public perception of the affected sector.In the following diagram, it can clearly be seen that the Public Administration and the Finance sectors suffered more seriously from damaged or unavailable systems, corrupted data files, or exfiltration of data compared to the other sectors:Figure 2 ENISA: Digital Impact by SectorSecure Element is a technical solution for digital payments via credit cards and mobile devices, as well as for identification and biometric purposes, such as passports and personal IDs.Since this device secures critical data, governmental bodies and private entities, such as credit card organizations, EMVCo[2], mandate that it be certified to Common Criteria EAL 5+ when using PP0084 – Security IC Platform Protection Profile with Augmentation Packages (Eurosmart, 2014)[3]. Till today more than 250 product certifications were done claiming for this PP.With the integration of the Secure Element in SoC, new challenges/threats were raised on top of the existing challenges/threats of the secure device with high resistance to physical and logical attacks:• Preventing the insecure state of the product by disturbing the boot process and enabling manipulation of the product by hostile software or malicious code.• Preventing content abuse of the data and code stored at the external non-volatile\volatile memory which is part of the SoC architecture by the attacker which accesses the external memory for disclosing or modifying the content of the external memory used by the secure component and by compromising confidentiality and/or integrity of secure content to be protected by the secure component.• Preventing Cloning of the content stored in the external memory or physical replacement of the external memory of the data and code stored at the external non-volatile\volatile memory.[2] EMVCo - https://www.emvco.com/about-us/overview-of-emvco/[3] Security IC Platform Protection Profile with Augmentation Packages : https://www.commoncriteriaportal.org/files/ppfiles/pp0084b_pdf.pdf• Preventing the ability of replay commands, the write, erase or responses to the read commands between the security component and the external memory, to affect the freshness of the content read from or written to the external memory. Preventing Unauthorized rollback of content.• Preventing the attempt to read the content of the external memory, record it, and later write it back to the external memory after the original content was updated by the Security component.• For SoC architecture that uses Secure Memory, the interface between the secure memory and the secure component should be protected from being blocked or intercepted by an attacker eavesdropping on the interconnection bus (e.g., by a man-in-the-middle attack), to disclose the user data and/or code data being written to or read from the secure external memory before security services are executed or finalized by the secure external memory.SoCs with integrated security functions appeared in the market and the security evaluation was done in a way of a mixture of PP0084 or part of it with extended requirements which might reflect the newly innovated device. No unified requirement. The challenge was to define all aspects of using and protecting the security functions when it is being integrated into the SoCThe methodEurosmart took the challenge and established a technical working group under its domain, ITSC. The subgroup includes Eurosmart members from the industry: semiconductor companies, software companies, ITSEF involved in evaluating security devices, Certification bodies, and consultants in this field.The national certification bodies were invited to the working group even though they are not Eurosmart members.On top of it, a liaisons\sharing was established with stakeholders who are referring \ interested \using this Protection Profile:A. Peers working groups: JHAS and ISCI-WG1.B. Organizations that reference the PP: FIDO, GlobalPlatform, GSMA.C. ENISA – for the alignment with CSA-EUCC which will be the scheme for this PP once the act will be implemented.The resultPP0117, Secure Sub-System in System-on-Chip (3S in SoC) Protection Profile includes the following:The TOE (Target of Evaluation) is "a Secure Sub-System (3S) implemented as a functional block of a System on Chip (SoC). The TOE implements a processing unit, security components, I/O ports and memories to provide a range of security functionalities covering a defined set of security objectives. The TOE provides its security features and security services isolated from the remaining SoC components, based on physical and/or logical isolation mechanisms. The TOE may rely on external memories to store content (data, code or both)."Figure 3: The Target of Evaluation (TOE)The TOE can be delivered as hard macro and/or programable macro, PL macro, as was defined in the team objectives.On top of it, the usage of external memory in different stages of the life cycle should be considered as well.The team strives to develop as generic as possible life cycle and highlight the new aspects of this architecture. It was clear that the new life cycle requires elaboration. With the cooperation with ISCI-WG1 a supplement guidance document, "Life-Cycle Model (LCM) Related Evaluation Aspects" was developed with more explanations related to the aspects that need to be fulfilled and assessed in the different phases of the life cycle.Figure 4: TOE Life CycleThe Protection Profile was structured with a base package of minimum requirements for any Secure Sub-System in a SoC, plus optional packages to address additional industry-specific needs arising from the architecture:• External Memory packages (Passive and Secure, volatile and non-volatile memory) – The restrictions related to the security of the data and code that are stored in the external memory.• Loader Package – The restrictions in loading functionality of the TOE Software or Composite Software from external memory.• Crypto Package - Framework for the integration of various cryptographic algorithms supported by the TOE. For addressing the need to be a generalized PP, this package, contrary to PP0084, doesn't define specific algorithms to implement but general instructions regarding the usage of recognized cryptographic algorithms.• Composite Software Isolation Package - The isolation features enable the separation between different software packages which may be delivered by different developers.Figure 5 PP Packages structureThe Security Problem Definition (SPD) which includes the assets to be protected, the threats, policies, and assumptions was developed in light of the collaboration with the JHAS group.In the Security Objectives section, dedicated objectives were defined related to the new approach of the TOE form (hardmacro\ PL macro).The base package of the Security Functional Requirements (SFRs) includes the PP0084 SFRs but for fulfilling the TOE need to be a Root of Trust, additional requirements for unique identification were included.The integration of the security sub-system in a non-secure SoC leads to the need to define the TOE as a way it provides its services isolated from the other SoC components based on physical and/or logical isolation mechanisms.The challenge in enabling integration of certified sub-systems in a non-secure system required new practices to be done by the developer and to be assessed by the ITSEF – the developer should instruct in which conditions the integration should be done and the ITSEF should verify that the integration was followed and no compromising of security was inspected during this process.Dedicated refinements related to the integration were added to the Security Assurance Requirements (SARs) for the ITSEF to verify the process was defined and done with no compromises.The evaluation was done by SGS with the supervision of BSI.SummaryPP0117 represents a significant advancement in cybersecurity certification for integrated systems. By providing a unified, flexible framework, it bridges the gap between traditional discrete certifications and modern integrated solutions, ensuring robust protection for sensitive data in an increasingly interconnected world.Winbond supports PP0117 by offering the W75F Secure Memory, which fulfilled the Secure External memory package. With Winbond EAL 5+ certified secure Flash, PP0117 can be claim in a composition with Winbond device and offer trusted external memory solution within SoC architectures. For more information, please visit Winbond website or download the latest Hardware Security White Paper.

As criminal organizations expand their use of information technology, the frequency of global financial fraud rises to historic levels each year, posing significant challenges for governments and the financial system. According to a Nasdaq Verafin research report, global financial and bank crime generated US$485.6 billion in losses in 2023, with total unlawful money transfers totaling US$3.1 trillion.The "Otter Coding" team discovered during the "2025 AI Wave: Taiwan Generative AI Applications Hackathon" that conventional financial institutions frequently employ manual review and outmoded model screening to prevent such incidents from occurring. This method is both time-consuming and difficult to address in a timely manner. The team proposed a financial fraud detection system that enhances financial institutions' anti-fraud response capabilities by integrating AI and other advanced technologies.The "Otter Coding" team ultimately triumphed in the highly competitive "Financial Innovation" group of Taishin Financial Holding.Otter Coding claims victory in the Financial Innovation category with an AI-driven fraud detection system built to outsmart global financial crime. Credit: CompanyAI-powered anti-fraud detection targets suspicious accountsThe "Otter Coding" team initiated a two-pronged approach to address the challenges associated with conventional financial fraud prevention. The first step is to employ the bank transaction records and fundamental account information supplied by Taishin Financial Holding to implement AI models for the identification of potential warning accounts. These accounts should be swiftly designated as warning accounts to enhance anti-fraud awareness and alleviate the staffing strain. The secondary purpose is to perform a comprehensive reverse analysis of the existing detection system utilizing large-scale language models to pinpoint possible areas for model enhancement. This will facilitate the ongoing development and enhancement of the anti-fraud system's accuracy.Following the completion of the Generative AI Workshop offered by Amazon Web Services (AWS) and the analysis of the data content provided by Taishin Financial Holding, the "Otter Coding" team members utilized their after-work hours a week prior to the official competition to convene and contemplate the direction of AI model design. They proposed the concepts of data-driven, model optimization, and automation, and ultimately resolved to design and construct the model from five major stages.During the model training phase, the team employed Amazon SageMaker Data Wrangler to integrate transaction records, essential account information, and alert data, perform feature engineering, and train and optimize the XGBoost classification model to achieve high precision and recall. Secondly, in the fraud prediction phase, transaction data is converted into feature vectors and input into the model for risk assessment, enabling the swift identification of high-risk accounts.During the third phase of result evaluation, the "Otter Coding" team computed Precision, Recall, and F1-score by juxtaposing the predictions with the actual list of fraudulent accounts and performed a comprehensive study of the error types.In the fourth round of AI analysis, the team utilized Anthropic Claude 3.5 Sonnet v2 model on Amazon Bedrock, a fully managed service that makes high-performing foundation models, to conduct second-level risk analysis, provide model optimization recommendations, and assist financial specialists in identifying deception patterns.Ultimately, the team employed AWS Amplify, the frontend development tool, to deploy the frontend UI and stored the data in Amazon S3, resulting in an immediate and user-friendly visual analysis interface, during the cloud deployment and presentation phase."Otter Coding" commended the comprehensiveness of both AWS's Generative AI Workshop and DIGITIMES' event planning and venue preparation. Building on this success, the team aims to achieve similar outstanding results in future competitions.

Telemarketing, a conventional sales channel, faces substantial challenges, such as a persistent attrition of qualified personnel, elevated labor costs, variable sales quality, and obstacles in scaling operations, despite the widespread adoption of digital transformation in numerous industries.At the second "2025 AI Wave: Taiwan Generative AI Applications Hackathon," the "Cloud Science Network Management" team presented an innovative solution that merges AI with user-centric design to transform telemarketing. This groundbreaking approach won the Eastern Home Shopping & Leisure's "Next-Gen Retail" competition.The "Cloud Science Network Management" team comprises individuals from the departments of Electronics, Computer Science and Engineering, Information Management, and Electrical Engineering at National Yunlin University of Science and Technology in Taiwan. Break away from conventional thinking, they leverage their professional expertise to develop an intelligent sales system addressing market challenges. This system uses AI technology to resolve long-standing industry pain points, while maintaining a humanized interactive experience.The team discovered that traditional telemarketing is confronted with three significant challenges following extensive market research. Initially, corporate profits are significantly impacted by the high costs of labor and training. Secondly, the performance levels of salesmen are inconsistent, and it is exceedingly challenging to standardize sales scripts. Thirdly, personnel recruitment bottlenecks that severely limit sales expansion.Although telemarketing continues to be a significant promotional channel, market research indicates that consumers are evidently opposed to interactions that appear to be overly mechanical. Consequently, the optimal solution must strike a balance between the efficacy of AI and the humanization of interaction in order to establish a sales process that is both strategic and highly consistent, while also ensuring consumers cannot detect they are interacting with an AI system rather than a human representative.To address these challenges, the team developed the "AI Smart Salesperson," a sophisticated solution employing AI technology for telemarketing. The system transforms AI into a highly effective telephone sales representative by incorporating data-driven analysis, real-time emotional assessment, and natural voice interaction technology. The operational workflow consists of four phases: first, inputting basic customer and product information; second, developing personalized sales strategies based on customer backgrounds; third, initiating the calling process where the customer's voice is transcribed using speech-to-text (STT) technology, dynamic responses are generated through a dialogue large language model (LLM), and these responses are converted into natural speech using text-to-speech (TTS) technology; and finally, using an analytical LLM to simultaneously assess customer emotions and responses, enabling real-time adjustments to sales strategy based on immediate feedback.The "AI Smart Salesperson" substantially reduces labor and training expenses while ensuring uniformity and consistency in sales methodologies, significantly enhancing both sales success rate and customer experience. The system's technical architecture is built on Amazon Web Services (AWS) and leverages multiple innovative technologies to ensure reliability and natural interaction. The core technologies include large language models responsible for dialogue generation, strategy planning, and real-time analysis, along with bidirectional text-to-speech (TTS) and speech-to-text (STT) conversion capabilities.For the AWS implementation, the team utilized Amazon Bedrock, a fully managed service that makes high-performing foundation models, to integrate models such as Anthropic Claude for dialogue generation. They employed Amazon SageMaker for machine learning model training and inference, and Amazon Transcribe for automatic speech recognition to enable speech-to-text functionality. The solution's data storage uses MySQL databases, while Amazon CloudFront handles web request delivery. The entire infrastructure runs on Amazon Elastic Compute Cloud (Amazon EC2) instances, providing the scalability and reliability needed for production deployments.This groundbreaking approach received unanimous recognition from judges in Eastern Home Shopping & Leisure's "Next-Gen Retail" category. The system features natural voice simulation for a genuine interactive experience indistinguishable from human interaction, standardized speech, and dynamic adjustment capabilities to maintain consistent and adaptable sales quality, along with a data-driven recommendation mechanism that significantly enhances the likelihood of a successful transaction.Furthermore, customers can promptly activate the platform by importing product information and listings, without the need to construct additional complex infrastructure, as a result of the out-of-the-box design. These benefits not only resolve the telemarketing dilemma that contemporary enterprises encounter, but they also establish a new paradigm for the application of AI throughout the entire industry.

Companies have begun to incorporate generative AI models with CAD drawings, design specifications, simulation data, and other information in order to improve design efficiency and accelerate innovation in response to the increasing competition in the commercial sector and volatile consumer demand. This facilitates the rapid generation of initial design drawings or concepts, reduces the length of product development timelines, and maintains the brand's essence.At the "2025 AI Wave: Taiwan Generative AI Applications Hackathon," "Famous on the Board" (a team name referencing PC hardware culture)unveiled an AI-assisted PC case design platform that addresses user pain points to achieve three objectives: "expediting the design cycle," "maintaining brand consistency," and "enhancing innovation efficiency and diversity of inspiration." This innovative platform combines image generation technology and natural language processing to convert user-defined requirements into structured specifications. This encompasses color, style, shape, material, ventilation, lighting, features, environment, and more, and it delivers high-quality rendered images in real time.The platform reduces communication disparities within the development team, enhances interactivity through real-time previews, enables version management for process traceability and restoration, and supports multiple models to increase rendering flexibility, thereby further alleviating the workload. Furthermore, it received recognition from the jury of Cooler Master's "PC DIY Innovation" category and was granted the Excellence Award.Famous on the Board wins Excellence Award with their AI-powered PC case design platform that blends speed, creativity, and brand identity. Credit: CompanyDriving inspiration diversity and enhancing innovation efficiencyThe "Famous on the Board" solution is distinguished by its ability to automatically document each attribute modification and rendering outcome, generate a visual design process timeline, and allow users to generate a PC case design from inception through the inspiration generation module. This allows users to scrutinize and trace discrepancies among multiple versions, enabling them to immediately revert to any historical version for iteration and optimization with a single click.The design of platform functions prioritizes fundamental features such as inspiration discovery, market trend analysis, design prompt generating, version control, and project collaboration management. A machine learning-generated database includes six key attribute databases essential for PC case design: design style, case material, color style, distinctive design, presentation form, and background atmosphere, helping designers rapidly produce design drafts that align with brand positioning.The AI-assisted PC case design platform employs Amazon Nova Pro, Anthropic Claude 3.7 Sonnet, Amazon Nova Canvas, and Stability AI's Stable Image Core on Amazon Bedrock, a fully managed service that makes high-performing foundation models from Amazon Web Services (AWS). It simultaneously supports natural language processing and creative image generation capabilities, thereby augmenting the intelligence and automation of the design process. React and TypeScript are employed in front-end development, complemented by Tailwind CSS to build a responsive user interface that supports internationalization (i18n). The backend architecture utilizes FastAPI and PostgreSQL 17.4 to guarantee data processing efficiency and reliability, while accommodating extensive asynchronous workloads."Famous on the Board" concluded that the essence of its success is rooted in the team's diverse composition. The team comprises development engineers with technical expertise and individuals from various sectors, particularly the gaming industry, each providing distinct perspectives on the application and conceptualization of AI. The challenging work was completed in under 30 hours through collaborative brainstorming and idea exchange, garnering unanimous recognition from the jury for the team's technical expertise, creativity, and degree of completion.

Smart manufacturing has become a global consensus in the manufacturing industry due to the rapid changes in consumer market demand. The implementation scenarios that incorporate generative AI technology are anticipated to further enhance the efficiency of production lines.At the "2025 AI Wave: Taiwan Generative AI Applications Hackathon," the victor of Chang Chun Group's "Smart Petrochemical Manufacturing" category, "Cloud Ninjas," adeptly developed an intelligent assistant system capable of promptly processing natural voice commands from on-site operators by integrating three key technologies: speech recognition, semantic comprehension, and task scheduling. This unique approach digitizes important operational experience and significantly enhances the natural fluency of human-machine communication via voice-driven operations.The "Cloud Ninjas" proposal's acknowledgment by the judges is ascribed to three primary features: "intuitive," "real-time," and "expandable." This transformation of human-computer interaction from a conventional one-way operation to a reciprocal comprehension and immediate response is facilitated by these attributes, which render the machine that "understands," "delivers," and "learns rapidly."The team meticulously defined a variety of command types, situational actions, and operational scenarios that were derived from real-world industrial contexts during the system design process. These scenarios included traversing from point A to point B, retrieving object A, activating button A, and articulating and reiterating content A. The scenario setting program simulates human operational modes through three state transitions: standby, pondering, and speaking. By employing the aforementioned rigorous designs, this AI machine assistant is capable of not only comprehending language but also autonomously executing the most appropriate action options based on the context.The "Cloud Ninjas", which was initially unfamiliar with one another, resolved to form a group to compete after attending a generative AI workshop from Amazon Web Services (AWS). They specifically targeted the manufacturing industry, which was more familiar to them. Despite the fact that all members possess a level of programming expertise, they are not acquainted with cloud tools, artificial intelligence, or other related technologies. Fortunately, they were able to demonstrate their results in the "2025 AI Wave: Taiwan Generative AI Applications Hackathon" after completing a series of pre-competition training courses from AWS and becoming more familiar with the use of a variety of tools.Throughout the development phase, "Cloud Ninjas" utilized several AWS generative AI services, including the generative AI-powered assistant, Amazon Q Developer to expedite program writing for software development, and Amazon Nova Pro, the multimodal model on Amazon Bedrock —a fully managed service that makes high-performing foundation models available—to attain precise semantic comprehension, response generation, and function calling. Simultaneously, it integrates the automatic speech recognition service, Amazon Transcribe, to convert speech to text, and the AI voice generator and text-to-speech tool, Amazon Polly, to provide natural and fluent vocal responses, thus creating an intelligent voice conversational robot."Cloud Ninjas" also integrates a real-time query response demo to autonomously search the Internet for user questions and return detailed information and related links in terms of generative AI technology. The team anticipates that these capabilities will expand to include high-value applications, such as immediate decision-making recommendations and operation manual searches.The global manufacturing industry is anticipating the attainment of more flexible operational objectives and increased efficiency through smart manufacturing as the world undergoes a tide of digital transformation. The "Cloud Ninjas" solution is a prime example of how generative AI is employed to develop AI assistants, serving as a key engine driving digital transformation in manufacturing.

In recent years, sporadic medical disagreements have arisen. Inadequate communication between the medical team and patients, together with their families, accounts for around one-third of medical mishaps, according to data from the Taiwan Patient Safety Net. Misunderstandings and distrust will ensue when a substantial information void exists between the attending physician and the anxiously expecting family members.During the second "2025 AI Wave: Taiwan Generative AI Applications Hackathon," the "Hero Song" team introduced a novel solution that integrates edge computing and AI technology, enhancing doctor-patient communication and securing victory in Advantech's "Intelligent Edge-Cloud Control" category.The members of the Hero Song team are from the Computer Science department at National Tsing Hua University in Taiwan. They discovered that the medical profession is confronted with two significant challenges following a comprehensive analysis. The initial issue is that there is a significant information divide between family members and physicians. The complete diagnosis and treatment instructions are not comprehended by family members, which leads to an increase in medical safety risks. The second reason is that family members frequently experience anxiety and the doctor-patient relationship is fraught as a result of their inability to comprehend the condition and the doctor's advice in real time. In order to resolve the aforementioned issues, the team developed "MediCAM," a cutting-edge medical communication system that integrates generative AI technology with Advantech's industrial AI camera (ICAM) and Amazon Web Services (AWS) cloud services.MediCAM comprises four principal functions: the first is "real-time image streaming," enabling family members to remotely observe patients' current conditions; the second is "automatic summary of medical advice," wherein the system records and converts doctors' instructions into comprehensible text; the third is "instant question and answer regarding patients' conditions," allowing family members to inquire about patients' statuses using natural language and receive informed responses based on actual medical data; the fourth is "AI daily patient status report," which autonomously generates a concise report on patients' conditions daily to assist family members in monitoring recovery progress. Not only can this system eradicate the information asymmetry gap, but it can also establish a foundation of mutual trust between doctors and patients, thereby significantly reducing medical risks caused by poor communication.The technical architecture of MediCAM is thoroughly designed, integrating cloud AI and edge computing.The team installed Advantech ICAM AI camera edge devices adjacent to patients' beds to collect on-site visual and auditory data. The system utilizes the Whisper model for voice-to-text and Amazon SageMaker, a machine learning service, to transcribe the doctor's instructions. Additionally, Yolo and FaceNet technologies are employed to verify the doctor's identity and perform facial recognition to ensure the legitimacy of the information source.The team employed Amazon Nova Pro on Amazon Bedrock, a fully managed service that makes high-performing foundation models, to process medical order summaries and enable intelligent question-and-answer capabilities. They used a retrieval-augmented generation (RAG) architecture to ensure the precision and dependability of the results. All data is encrypted and transmitted securely to protect patient confidentiality. AWS provides reliable support for streaming, access, and security control, enabling the consistent processing of large quantities of image, audio, and text data.MediCAM, which is both practical and inventive, was the standout technology proposition in the Advantech technology proposition category for the "Hero Song" team.The jury stated that this system possesses numerous advantages: the high level of immediacy enables family members to comprehend the patient's condition promptly; the information is transparent and objective, while the automatic transcription and summarization of medical orders prevent information omissions or misinterpretations; the intuitive interactive experience permits users to obtain answers by merely posing questions in everyday language, significantly reducing the barrier to understanding. The solution utilizes AWS cloud services and can be rapidly scaled to multiple hospital campuses. This technology not only significantly diminishes the likelihood of ineffective communication between doctors and their patients, but also facilitates the integration of AI applications into Taiwan's healthcare system.