AI Solution Platform-EE Awards-EE Times

Product：Gen 4 SiC MOSFET Technology Platform

Recommended reasons：第四代 (GEN 4) 碳化矽MOSFET技術平台專為改善高功率設計中常見的開關行為與設計挑戰而開發。此平台提供750V、1200V及2300V等多種電壓等級，為Wolfspeed功率元件產品線制定了長期技術發展藍圖。第四代 (GEN 4) 碳化矽MOSFET技術將為高功率車用領域、工業應用及再生能源系統的設計者提供顯著的性能增強。專為極端環境而設計，確保長期可靠運作第四代 (GEN 4) 元件具備高達2.3微秒的短路耐受時間，為關鍵應用提供額外的安全餘裕。相較於前代技術，該平台將宇宙射線瞬態故障率（FIT）提升達100倍，確保在不同海拔高度下皆能維持可靠性能。體二極體設計優化可提升系統效能與耐用性，實現更快切換速度、更低損耗與減少振鈴現象——使VDS突波降低80%。精選的第四代晶片通過185°C連續運作與200°C有限壽命運作認證，為追求極致性能的設計提供彈性空間。專為全面提升系統效率而設計第四代 (GEN 4) 元件可在工作溫度下實現特定導通電阻降低達21%，有效減少輕載與滿載時的傳導損耗。在硬開關應用中，更能帶來開關損耗降低15%的優勢。此效能提升佐以第四代元件優異的175°C RDSON表現，充分體現Wolfspeed對實際工作環境中可靠性能的承諾。專為降低系統成本與縮短開發週期而設計新一代平台能採用更小型、更經濟的被動與濾波元件，從而縮短系統開發時間並降低成本。此外，第四代 (GEN 4) 元件可在相同封裝尺寸下實現高達30%的功率輸出提升。其整合的新型軟體二極體設計能大幅降低反向恢復時的EMI干擾，不僅簡化EMI認證流程，更可縮減EMI濾波器的尺寸需求。

More Vote

Footprintku

Product：footprintku.ai

Recommended reasons： 1. 結合DaaS、SaaS與ODaaS架構，提供符合DFM要求的ECAD Library共用平臺。 2. 具備高度客製化彈性，滿足不同規模企業的多元設計需求，實現設計資料共用。 3. 提供數位化工具與資料服務，降低企業數位化導入門檻，加速開發流程，強化市場反應與競爭力平臺功能: DFM Assistant-打造DFM自動化引擎 Download ECAD-一鍵取得符合DFM的ECAD Library Request ECAD-全面支援新零件的ECAD Library ECAD Studio-化繁為簡輕鬆管理ECAD Library

More Vote

STMicroelectronics

Product：5.5kW advantages of hybrid TCM/CCM control in interleaved TTP PFC Reference Design based on STM32G4 Digital Controller

Recommended reasons：該解決方案通過結合ST的高性能650V、40 mΩ矽碳化物（SiC）MOSFET以及封裝在TOLL封裝中的600V、30 mΩ高壓MOSFET，達到了超過99.2%的卓越峰值效率。這一高效率通過優化的柵極驅動電路得到了進一步增強，該電路採用了ST的電隔離柵驅動技術（STGAP），提供了精確的開關操作並最小化損耗。該系統由STM32G474微控制器全面數位化控制，支援混合信號處理，以實現精確和靈活的電源管理。在交錯配置下運行，該設計顯著降低了脈動電流，並允許更小的輸出電容。在半負載以下，系統運行在時脈調變模式（TCM），實現零電壓開關（ZVS），以提高效率。在半負載以上，它切換到連續導通模式（CCM），減少電感電流脈動和導通損耗。在滿負載下，功率因數保持在0.98以上，總諧波失真電流（iTHD）低於5%，符合嚴格的電能品質標準。此設計還採用了一種創新的2合1電感，將雙電感整合到一個單一的磁性元件中。這樣可以減少PCB面積和元件數量，使電源階段更加緊湊且具成本效益。改進的磁耦合降低核心損耗和漣波電流，減輕輸出電容器的壓力，增加整體系統的效率和可靠性。此外，2合1電感改善了熱管理，支援在高負載下的穩定運行，並延長系統的壽命。此外，該設計在功率因數校正（PFC）階段採用了STPM32計量解決方案，提供全面的電力監控功能。數位控制策略確保在208 Vac至277 Vac的輸入電壓以及10%至100%的負載條件下具有高效率。在10%負載時，效率保持在97.88%以上，而在滿載時則超過98.46%。主要特點包括寬輸入電壓範圍（208-277 Vac）、高開關頻率操作（50 kHz至180 kHz）以及強大的功率因數校正。這一參考設計代表了一種高效率、靈活且可靠的電源解決方案，十分適合高要求的伺服器電源應用。

More Vote

Arm

Product：Arm Kleidi (KleidiAI and KleidiCV)

Recommended reasons： 1. 具彈性的各種核心組合 Arm Kleidi 提供具有彈性的各種核心組合，協助在框架上強化人工智慧。在 Arm 架構上提供廣泛的多面向人工智慧進展成果，從實現更多人工智慧功能、提升人工智慧準確性、到加速或減少記憶體負擔。 2. 精簡、高效且輕量全新 KleidiAI 及 KleidiCV 效能函式庫非常輕量精簡，沒有記憶體配置，也沒有其他函式庫相依性或二進位版本，因此更容易採用，並能快速高效地整合至現有框架程式碼基底。 3. 釋放大眾市場的人工智慧效能 Kleidi 可於框架層級最佳化人工智慧，而每項最佳化都能讓數十億個 Arm 架構裝置的數百項工作負載從中獲益。應用程式開發人員只要在 Kleidi 最佳化框架執行模型，就能依預設達成頂尖效能。 4. 實現大規模的生成式人工智慧 Kleidi 可協助以最輕鬆快速的方式，讓作業最繁重的人工智慧推論工作負載部署於 Arm。KleidiAI 函式庫可在爆炸性成長的生成式人工智慧及大型語言模型 (LLM) 市場中導入同級最佳效能，部署範圍涵蓋雲端資料中心至邊緣的受限裝置。 5. 讓所有人隨處享有最佳化人工智慧 Kleidi 的目標是在全系列的 Arm Neoverse 及 Arm Cortex-A CPU 產品中，從雲端至邊緣輕鬆地促成最佳化。效能函式庫利用特定技術在 Arm 架構強化人工智慧功能，例如 Arm Neon、Arm 可擴展向量延伸指令集 (SVE) 及 Arm 可擴展矩陣延伸指令集 (SME)。 6. 降低開發人員所需費用 Arm Kleidi 函式庫的願景是將其直接整合至關鍵人工智慧框架中，包括 MediaPipe (透過 XNNPACK)、LLAMA.ccp、PyTorch (透過 ExecuTorch) 及 TensorFlow Lite (透過 XNNPACK)。整合後，開發人員就能自動享有 Kleidi 最佳化框架的效能強化效益，無須承受任何直接費用。

More Vote

DeepMentor Inc.

Product：DeepExpert

Recommended reasons： DeepExpert is an enterprise-grade GenAI deployment platform tailored for secure, on-premise, and domain-specific language model applications. It offers a modular no-code toolchain that enables enterprises to customize, fine-tune, and deploy LLMs or vision models without requiring in-house AI experts or public cloud resources. This platform empowers businesses and government agencies to swiftly create a proprietary AI intelligent system tailored to their enterprise needs. We ensure secure data operations within client's environments for confident AI transformation without security compromises. The platform consists of three key modules: o Data Expert: Supports custom training data augmentation and knowledge distillation for private AI system development. Assists in structured data generation and annotation to accelerate supervised learning. o Finetune Expert: Provides visual, no-code interfaces for domain-specific LLM customization, with built-in training progress monitoring (loss curves, accuracy charts, inference validating). o Document Expert: Automatically extracts, summarizes, and enables semantic search across enterprise documents, regulations, and technical manuals. It handles complex entities like tables and images, and supports fact-checking with intuitive UI, enabling precise knowledge retrieval and reasoning. DeepExpert integrates DeepMentor's end-to-end AI ecosystem: • Underlying AI IP: While an application platform, DeepExpert leverages the efficiency and miniaturization capabilities enabled by DeepMentor's DeepTransformCore and DeepLogCore IPs for optimal performance on underlying hardware, including the Mentor series workstations and servers. • Knowledge Distillation: Utilizes techniques like knowledge distillation to create efficient sLLMs that retain the decision accuracy of larger models but are optimized for local deployment. • Developer-Friendly Tools & APIs: It provides a comprehensive set of Developer Tools & RESTful APIs for inference and RAG. This includes features for document ingestion and database connection. • Full-Stack Integration: DeepExpert is a testament to DeepMentor's ability to seamlessly integrate AI algorithms, chip design, hardware, firmware, and services, offering a truly one-stop solution for deploying robust GenAI applications.

More Vote

DAINVI Limited

Product：MonoSight™ AI

Recommended reasons： MonoSight™ is a breakthrough AI vision system purpose-built for high-mix, low-volume (HMLV) manufacturing, where traditional AI struggles to deliver value due to data scarcity and deployment complexity. Conventional AI requires extensive data collection, annotation, model training, and iteration—an expensive and time-consuming process that rarely makes commercial sense in fast-changing production environments. In contrast, MonoSight™ uses self-supervised visual learning and a training-free, patch-level anomaly detection engine to deliver accurate results from just a single reference image. This enables manufacturers to: • Deploy instantly without labelled data • Detect rare or unknown defects from day one • Scale AI adoption across diverse product lines with minimal effort MonoSight™ empowers teams to move beyond proof-of-concept and into real-world impact, unlocking the value of AI in places it was previously impractical. It’s not just a smarter way to see—it’s a smarter way to scale. MonoSight™ – All it takes is one

More Vote

YUAN High-Tech Development Co., Ltd.

Product：EYE6N0-S AI Smart Camera

Recommended reasons： In dynamic public environments, real-time awareness of crowd movement and safety is essential for effective site management. To meet this need, YUAN introduces the EYE6N0 Series AI Cameras, an integrated smart solution powered by NVIDIA^® Jetson Orin™ NX Super, and seamlessly working with SmartFlow, SmartNVR, and SmartVMS to form a complete AI vision system. ➤ Flexible Zoom & Dual-Lens Coverage EYE6N0 offers auto-focus and auto-iris capabilities, with 3x optical zoom in the standard model and up to 25x in the PRO version. Dual-lens design version captures both near and far views simultaneously, enhancing scene flexibility and detail. ➤ All-Weather High-Quality Imaging With HDR, infrared illumination, and IP66-rated weather resistance, the EYE6N0 ensures clear images across day and night, in both indoor and harsh outdoor conditions. ➤ Custom AI Logic with NexVDO SDK The built-in NexVDO SDK supports up to 24 AI models, allowing users to define custom monitoring logic and detection rules across various deployment scenarios. Integrated with SmartFlow, the system enables real-time visitor tracking, area-based headcounts, and average dwell time analysis. Heatmaps and dashboards support instant operational adjustments—ideal for retail, exhibitions, and transit hubs. SmartNVR connects video streams to backend storage and analytics, with plug-and-play AI modules for person, vehicle, face, and license plate detection. The interface is intuitive, supporting event triggers and automated alerts to streamline safety management. SmartVMS, a Linux-based cloud platform, centralizes NVR resources across locations. Through a modular web dashboard, operators can monitor live events and statistics in real time, with flexible master-slave architecture, public/private cloud deployment, and mobile access. From image capture to AI analytics and cloud-based management, YUAN delivers a fully integrated AI monitoring solution, empowering users with real-time situational awareness and enhanced security control.

More Vote

Taiwan AI Service Corporation

Product：AFS POD Service

Recommended reasons： TASC’s AFS (AI Foundry Service) POD Service is a turnkey AI infrastructure solution that seamlessly integrates servers, data center systems, and AI-optimized software into a pre-validated package. Designed for governments, enterprises, and global GPU market, it supports the full spectrum of AI data center demands—from building, operating, training, to deployment—at scale. The solution enables rapid deployment of on-premises AI data centers in as little as 8 weeks, backed by enterprise-grade SLAs. It features advanced resource management, workflow orchestration, and billing system integration, delivering end-to-end operational efficiency. With over 7 years of AIHPC expertise and 160+ PetaFLOPS delivered, TASC’s solution empowers organizations to scale generative AI and HPC workloads while reducing cost, power, and complexity. It provides governments with a secure, sovereign AI infrastructure, enabling enterprises to build autonomous AI capabilities and accelerate national AI strategies, thereby driving industrial innovation.

More Vote

EmTDLab Space Division - Emerging Technologies Development Laboratory

Product：Systematic Materials Discovery Engine - Symade

Recommended reasons： Symade.ai is a proprietary materials discovery engine that search for novel radiation shielding materials having also strong mechanical properties. EmTDLab has developed a patented Materials Informatics process to discover and produce a generation of entirely new radiation shielding materials primarily aimed at satellites, spacecraft, and space vehicles. The most critical capabilities are the systematic search and optimization for new radiation shielding materials using a proprietary evolutionary algorithm and its ability to accurately estimate materials’ density and stability and predict their radiation shielding properties. The same algorithmic process can be tuned to fulfil the needs of other industries as the development of such novel materials with a pure algorithmic approach will result in an expansion of materials design capabilities. As of now, 45 new alloy materials have been discovered and are ready for testing and commercialization.

More Vote

NXP Semiconductors

Product：NXP eIQ^® Time Series Studio

Recommended reasons： To advance AI on the edge, we are introducing the eIQ^® Time Series Studio (eIQ TSS), a new tool in our eIQ AI and machine learning development software family. eIQ TSS features an automated machine-learning workflow that streamlines the development and deployment of time series-based machine learning models across microcontroller (MCU) class devices such as the MCX portfolio of MCUs and i.MX RT portfolio of crossover MCUs. Time Series Studio supports a wide range of sensor input signals, including voltage, current, temperature, vibration, pressure, sound, time of flight, among others, as well as combinations of these for multimodal sensor fusion. The automatic machine learning capability enables developers to extract meaningful insights from raw time-sequential data and quickly build AI models tailored to meet accuracy, RAM and storage criteria for microcontrollers. The tool offers a comprehensive development environment, including data curation, visualization and analysis, as well as model autogeneration, optimization, emulation and deployment.

More Vote

SUNIX CO.,LTD.

Product：NPU AI Accelerator Card

Recommended reasons： SUNIX AIEH4000 — Unique Design, Powerful Features, Smarter Edge AI Unique Design The AIEH4000 features a low-profile, single-slot PCIe x16 form factor, allowing seamless integration into existing systems such as IPCs, NVRs, and edge servers without major modifications. Its modular and compact design enables flexible deployment and AI upgrades on legacy hardware platforms. Special Features Powered by four Hailo-8™ AI processors, the card delivers up to 104 TOPS of AI inference performance. Designed for low power consumption, it supports real-time processing of multiple video streams—ideal for traffic monitoring, smart manufacturing, and security surveillance. Built-in thermal and power management ensures stable operation in demanding environments. Function and Technology Used AIEH4000 supports leading AI frameworks (TensorFlow, ONNX) and comes with Hailo’s SDK for easy model deployment. It achieves exceptional results in benchmarks—ResNet-50 at 5,328 FPS and YOLOv5m at 872 FPS. With high-throughput data handling and parallel processing architecture, it significantly offloads the CPU and accelerates overall system performance. The AIEH4000 is more than just an accelerator—it is a scalable, high-efficiency AI solution built for demanding edge applications where performance, size, and power efficiency all matter.

More Vote

Alif Semiconductor

Product：Ensemble New Generation EdgeAI MCU with Crtex-M55 and Ethos-U85

Recommended reasons： The new E4, E6 and E8 Ensemble MCUs and fusion processors integrate the latest neural processing unit (NPU) from Arm^®, the Ethos ™ -U85, with Alif's unique high bandwidth memory and intelligent power management architecture to support NPU operators required by transformer networks to efficiently execute generative AI workloads. Transformer networks enrich endpoint use cases by focusing on the most relevant parts of older portions of an acquired input sequence while rapidly processing in parallel the current acquired sequence to enable deeper contextualization. For localized use this results in an instant, accurate, and more satisfying experience when using multimodal AI in endpoint products focused on vision, voice, text, and sensor fusion without relying on the cloud. The new family introduces a set of improvements to give even higher AI/ML performance at lower power consumption, and to provide more integrated support for embedded vision and audio applications. New features integrates a wider internal memory bus topology bringing abundant localized data bandwidth for NPU consumption, dual MIPI-CSI camera interfaces, a fast hardware image signal processor operating up to 200 frames per second, and a pair of faster interfaces to external memory each capable of 800MBytes/s. And finally, new broader options for SRAM data retention gives designers more flexibility to wake from sleep faster.

More Vote

AICE Inc.

Product：Marketing Deep Dynamic AI Agent System：ADGo

Recommended reasons： ADGo is the world’s first and only production-ready AI system for creative strategy and marketing ideation. Powered by MAGICC (Multi-Agent Generative IDEA Co-Creation), ADGo simulates a full-stack marketing team through 60+ AI agents trained in strategic reasoning and campaign execution. It helps marketers generate 15,000-word marketing proposals in just 15 minutes—with built-in support for Chinese, English, Japanese, and five APAC market data sources. No technical skills required. From research to creative content, ADGo streamlines the entire process into one seamless workflow. In just six months, it has been adopted by 200+ enterprises, generated over 10,000 outputs, and consumed 1 billion tokens—achieving 72x faster creative production and enabling smarter, faster global marketing decisions.

More Vote

AMD

Product：AMD Instinct™ MI350 Series GPUs

Recommended reasons： AMD Instinct™ MI350 Series GPUs, featuring the cutting-edge new 4th Gen AMD CDNA™ architecture, are equipped with powerful and energy-efficient cores, up to 288GB of HBM3E memory, and 8TB/s bandwidth. By supporting next-generation FP6 and FP4 data types (including matrix sparsity), they redefine AI acceleration to achieve lightning-fast AI inference and training, setting new benchmarks for performance, efficiency, and scalability in Generative AI and High-Performance Computing (HPC). AMD Instinct MI350 Series includes the Instinct MI350X and MI355X GPUs and Instinct platforms, offering a 4x increase in AI compute capability and a 35x leap in inference performance compared to the previous generation, helping various industries achieve transformative AI solutions. The MI355X OAM 288GB delivers up to 2.1x AI performance, 2.1x HPC performance, and 1.6x memory capacity compared to the NVIDIA B200 SXM5 180GB. At the same time, the MI355X also offers significant improvements in price-performance, generating up to 40% more tokens per dollar compared to competitor solutions. AMD Instinct MI350 Series Platforms integrate 8 fully interconnected MI355X or MI350X GPU OAM modules onto an industry-standard OCP design via 4th Gen AMD Infinity Fabric™ technology, with an industry-leading 2.3TB HBM3E memory capacity for high-throughput AI processing. These ready-to-deploy platforms support a variety of systems, from standard air-cooled UBB-based servers to ultra-dense Direct Liquid Cooled (DLC) platforms, helping to accelerate time-to-market and reduce development costs when integrating AMD Instinct MI350 Series GPUs into existing AI rack and server infrastructures. Leading hyperscalers, model builders, and AI companies worldwide have widely deployed AMD Instinct GPUs, including Meta, OpenAI, Microsoft, Oracle Cloud Infrastructure (OCI), and xAI.

More Vote

MosChip Technologies

Product：MosChip DigitalSky GenAIoT Solution Suite

Recommended reasons： Accelerating intelligent product development often calls for reducing the time spent on developing foundational infrastructure - device connectivity, AI model tuning, automation, and cloud orchestration. The DigitalSky GenAIoT accelerator platform is designed to do exactly that. It provides a set of modular, pre-engineered components that help teams build, scale, and maintain connected ecosystems faster. The DigitalSky GenAIoT accelerator platform is built up for 4 different modules, designed to accelerate time to market and improve operational efficiency. - Integrated IoT & Connectivity Suite: This module supports secure device onboarding, edge gateway management, unified management, real-time insights, and adaptive security. This gives a robust starting point for real-time device-to-cloud operations. - Cognitive Intelligence Suite: This suite incorporates AI into the product workflow, enabling scalability and flexibility. With 75+ production-ready foundational AI models, which consists of 30+ Edge AI models, and 20 GenAI solutions covering speech, vision, anomaly detection, and decision logic, all equipped to enable informed decision making, achieve greater operational efficiency, and accelerate AI model development. - Unified Automation Suite: Our 10+ automation agents help engineering and QA teams scale efficiently - with pre-built automation agents supporting RPA, regression, test automation, and provisioning. It assists in streamlining workflows via autonomous agents and dynamic orchestration, thereby reducing operational bottlenecks. - Digital-Native Suite: It is designed to guide enterprises through cloud-native transformation, microservices modernization, and integrated data intelligence. The DigitalSky GenAIoT Suite is designed to address dynamic requirements of various industries equipping them with a scalable and future-ready framework for developing next-generation smart products backed by cognitive intelligence.

More Vote

Nuvoton Technology Corporation

Product：NuMicro^® M55M1

Recommended reasons： The NuMicro^® M55M1 microcontroller series uniquely integrates a high-performance Arm^® Cortex^®-M55 core with an Arm^® Ethos™-U55 Neural Processing Unit (NPU), enabling real-time AI inference and digital signal processing directly on edge devices. Equipped with DSP extensions, vector processing, an FPU, and an NN accelerator, it supports comprehensive security features, including Secure Boot, TrustZone, encryption accelerators, and tamper detection. Additionally, the M55M1 provides extensive high-speed communication interfaces and precise analog control units, ideal for audio processing, machine learning, and industrial control applications, delivering a robust, secure embedded solution.

More Vote

AsicAI Co., Ltd.

Product：AsicAI™ Link

Recommended reasons： Beyond offering highly competitive NPU IP, AsicAI innovatively provides a one-stop AI solution platform, AsicAI™ Link, encompassing AI model design, IP development, and SoC integration. This service enables rapid integration of AI functionalities into chips, significantly boosting deployment efficiency and thus accelerating the widespread adoption and practical application of AI technology. The core of our one-stop service is a comprehensive and highly efficient suite of software tools and Application Programming Interfaces (APIs), designed to simplify the entire process of AI application development from concept to realization, offering a clear, efficient, and standardized workflow. Key features and advantages include: ．AI Model Hardware Friendliness Verification: Automatically analyzes the operations within AI models slated for deployment, ensuring they maximize the utilization of AsicAI's Neural Processing Unit (NPU IP). ．Support for Mainstream Model Frameworks: Supports popular frameworks like PyTorch, TensorFlow, and Caffe, eliminating the need for extensive model conversion time and enabling easy migration to hardware platforms. ．Ultimate Performance and Resource Optimization: Capable of model lightweighting and performance optimization, including critical model quantization techniques. This effectively reduces latency and shrinks model size while maintaining inference accuracy. By converting floating-point operations to more efficient integer operations, our empirical data shows that model size can be reduced by an average of 2-4 times, and inference speed can be boosted by 4-16 times. This significantly decreases power consumption and memory, creating faster and more energy-efficient AI applications. ．Customized Model Compilation: Offers two compilation methods for different application scenarios, balancing performance and versatility: General Compilation and Advanced Compilation. General Compilation does not require pre-compilation for specific hardware, allowing for real-time compilation and execution, suitable for general applications. Advanced Compilation is designed for end-user applications with higher demands for low latency or memory constraints, where AsicAI performs advanced and customized compilation to enable models to achieve superior performance and efficiency during inference. This service has successfully attracted several leading listed IC design companies for technical collaboration and solution adoption. Critically, two of these listed IC company clients have already successfully entered mass production phases with our solutions. Notably, an AI chip utilizing our technology for Microsoft's Human Presentation Detection application has garnered the attention and adoption of a leading international laptop manufacturer, with products anticipated to launch by the end of 2025. Furthermore, we've successfully secured a cutting-edge 7nm process AI chip design service project, providing comprehensive turnkey solutions ranging from NPU IP design and integration to complete System-on-Chip (SoC) design.

More Vote

INFINITIX INC.

Product：AI-Stack

Recommended reasons： AI-Stack, developed by INFINITIX, is an NVIDIA-certified Solution Advisor platform for efficient AI infrastructure management. It solves challenges in GPU utilization, deployment complexity, and multi-team resource coordination. With advanced GPU slicing, distributed training, and integrated with robust containerization and efficient MLOps workflows, AI-Stack reduces AI environment setup time to just 1 minute, it boosts GPU usage from 30% to 90%. AI-Stack is widely used across industries to accelerate AI adoption and digital transformation.

More Vote

DigWise Technology

Product：DTCO.GenAI (Virtual Silicon)

Recommended reasons： Generative Al GAN Diffusion Model WAT Super Resolution (WAT-SR) High-Efficiency SPICE-Silicon Bias Modeling (He-SSBM) High-Fidelity Generative Monte Approximation (HΣ-GMA) Multivariate Normal Distribution

More Vote

Data Systems Co., Ltd.

Product：Enterprise AI Assistant

Recommended reasons： Enterprise AI Assistant is an innovative application developed with generative AI technology, designed to provide enterprises with automated document image recognition and data entry services. Addressing the operational challenge of handling large volumes of paper-based documents, receipts, and vouchers, Enterprise AI Assistant significantly reduces manual data entry efforts, improves operational efficiency, and minimizes the risk of human error. The solution offers a range of key features and capabilities. First, its voucher image recognition function automatically identifies various types of documents, including invoices, receipts, and purchase orders. Second, the automated data extraction and entry feature accurately captures critical information from scanned images and inputs it directly into enterprise systems such as ERP and CRM. Additionally, the system includes data verification and correction mechanisms to ensure the accuracy and reliability of extracted data. To meet diverse business needs, Enterprise AI Assistant supports multiple image formats and offers custom integration services, enabling seamless connectivity with existing enterprise information systems. What sets Enterprise AI Assistant apart is its unique combination of generative AI models and optical character recognition (OCR) technology. Compared to traditional OCR, it can process more complex and irregular document formats, with the ability to continuously learn and adapt to new layouts and templates. In summary, Enterprise AI Assistant is a practical and innovative AI-powered application that delivers efficient and accurate document image recognition and data entry solutions. It has strong potential to make a significant impact in the field of intelligent enterprise transformation.

More Vote

Harbor Technology Solutions Co., Ltd.

Product：Three-layer AI architecture

Recommended reasons： Three-Layer AI Architecture: A Scalable and Efficient Platform for Real-Time, Self-Optimizing AI Applications The three-layer AI architecture distributes AI algorithm workloads across different layers, allocating computing resources based on each layer's functionality and hardware capabilities. This enables a high-performance, real-time, and continuously self-optimizing AI application platform. - Sensor AI: Our sensor features a compact AI model to perform basic data preprocessing and anomaly detection. This reduces unnecessary data transmission, easing network bandwidth and backend load. - Gateway AI: Centralize data from multiple sensors and process complex model inference. Raw data is also preserved for backend training and traceability. Supports parallel multi-tasking and flexible device scalability. - Server AI: Centrally manages all edge devices and AI model versions. Executes model retraining and automated deployment. Integrates data across devices and sites for high-level logic computation and decision-making. This architecture offers several technical advantages: - Layered computing design and task distribution - Large-scale real-time data processing - Flexible product development and AI model deployment - High system stability and maintenance efficiency - Continuous AI learning and automatic model updates Compared to traditional centralized AI systems, the three-layer architecture is better suited to handle high data volumes and multi-sensor environments. It enhances computational efficiency and system availability, making it ideal for industries such as smart manufacturing, industrial automation, smart cities, and IoT that demand real-time data processing. It is the optimal solution for enterprises aiming to achieve AI-driven transformation.

More Vote

Hyson Technology Inc.

Product：Al Image Demoiréing

Recommended reasons： HysonTech Image Demoiréing Technology uses AI to address the problem of unwanted moiré visual artifacts that cause poor image quality when using a mobile phone or camera to capture another digital screen. When the pixel arrangement of the display being photographed or the pattern of the object being photographed is inconsistent with the order of the photosensitive elements of the camera, the two arrangements will be superimposed into irregular colors and shapes, called moiré patterns and rainbow patterns. Unlike other image restoration tasks, the difficulty in eliminating moiré patterns is that they appear in a wide range of frequencies with irregular shapes and rainbow-like colors. Therefore, HysonTech has developed a multi-scale fusion neural network, which includes dilated-dense attention, multi-scale feature interaction, and multi-kernel strip pooing, which can capture moiré and rainbow patterns, and separate and extract the features of the real image to restore the moiré image, while enhancing the de-texturing effect. This technology has been used in the smart manufacturing processes of panel and chip manufacturers to significantly improve the defect detection performance of each manufacturer, and has been integrated into smartphones to solve the moiré problem generated during photography. Features: (1) Single Input Image Only, Auto AI-Demoireing for CCD/Phone Images. (2) Support 1080P/2K/4K/8K Full Resolution Image. (3) Support Both Cloud-AI Computing and Edge-AI Computing. (4) Add-on Solution: Software Upgrade Only, No Hardware Change.

More Vote

STMicroelectronics

Product：5.5kW advantages of hybrid TCM/CCM control in interleaved TTP PFC Reference Design based on STM32G4 Digital Controller

Recommended reasons： The solution achieves an outstanding peak efficiency exceeding 99.2% by integrating ST’s high-performance 650V, 40 mΩ Silicon Carbide (SiC) MOSFETs alongside 600V, 30 mΩ High Voltage MOSFETs housed in TOLL packages. This high efficiency is further enhanced by an optimized gate driving circuit featuring ST’s galvanic isolation gate driver technology (STGAP), which ensures precise switching and minimizes losses. The system is fully digitally controlled by the STM32G474 microcontroller, which supports mixed-signal processing for precise and flexible power management. Operating in an interleaved configuration, the design significantly reduces ripple current and allows for a smaller output capacitance. Below half load, the system runs in TCM mode, enabling Zero Voltage Switching (ZVS) to enhance efficiency. Above half load, it switches to CCM mode, reducing inductor current ripple and conduction losses. At full load, the power factor remains above 0.98 with total harmonic distortion current (iTHD) below 5%, meeting stringent power quality standards. This design also utilizes an innovative 2-in-1 inductor, integrating dual inductors into a single magnetic component. This reduces PCB area and component count, enabling a more compact and cost-effective power stage. The improved magnetic coupling lowers core losses and ripple current, reducing stress on output capacitors and enhancing overall system efficiency and reliability. Additionally, the 2-in-1 inductor improves thermal management, supporting stable operation under demanding loads and contributing to longer system lifetime. Additionally, the design incorporates the STPM32 metering solution on the PFC stage, providing comprehensive power monitoring capabilities. The digital control strategy ensures high efficiency across input voltages from 208 Vac to 277 Vac and load conditions ranging from 10% to 100%. Efficiency remains above 97.88% at 10% load and exceeds 98.46% at full load. Key features include a wide input voltage range (208-277 Vac), high switching frequency operation (50 kHz to 180 kHz), and robust power factor correction. This reference design represents a highly efficient, flexible, and reliable power solution ideal for demanding server power applications.

More Vote

Data-DI Solutions

Product：AltaBots.ai

Recommended reasons： AltaBots.ai stands out with its highly flexible, No Code Agent development platform, fundamentally transforming how businesses adopt AI. Its core strength lies in empowering business users, regardless of coding knowledge, to easily build AI applications, drastically cutting time from concept to execution and truly enabling them to "get started in minutes." A key feature is its extensive library of Agent templates and tools, offering over 100 ready-to-use Agent templates and 40 cross-industry, cross-departmental AI application templates covering core business scenarios like customer service, marketing, and operations. Uniquely, it boasts over 90 built-in AI tools for seamless integration, allowing AI to directly serve business needs and significantly reducing additional development efforts. Technologically, AltaBots.ai not only supports integration with all major LLMs, ensuring maximum flexibility for enterprises in choosing large language models. Furthermore, its innovative "AI Landing Coaching Program", which combines scenario consulting with application building and support from professional consultants, ensures AI truly delivers business value. This goes beyond mere tool sales, proving crucial for successful AI implementation.

More Vote