Into the era of education informatization 2.0 driven by AI, ESWIN Computing harnesses smart technology to reshape the underlying logic of teaching and learning.
Based on a terminal-edge-cloud integrated architecture, ESWIN Computing’s smart education solutions centered on high-performance edge computing devices are equipped with multimodal algorithm engines to analyze education data in real time, enable intelligent matching of resources, and enhance teacher-student interaction, covering the whole process including teaching, learning, evaluation, monitoring, testing, and management. By combining standard capabilities and personalized services, the solutions empower the digital transformation of education across all scenarios and provide technical support for bespoke instruction.
ESWIN Computing will further deepen the integration of smart edge and education and strive for technology breakthroughs in lightweight AIGC deployment and LLM edge inference for education, so as to form a smart closed loop of teaching, learning, evaluation, monitoring, testing, and management. The goal is to harness technology to contribute to individuals well-rounded development and jointly build a new smart education ecosystem.
ESWIN Computing’s edge computing devices for smart education use the EIC7700X, the company’s self-developed RISC-V edge computing SoC. Thanks to high performance, low power consumption, and other technical strengths, multimodal AI algorithms in education settings are integrated to locally collect audio/video data and conduct real-time analysis in milliseconds. Through intelligent processing at the edge, this ensures timely data processing and the security of education data, reducing the data transmission latency and lowering the risk of privacy leakage.
ESWIN Computing’s Open Standard Module (OSM) is an open standard-based, modular computing platform that complies with the dimension, interface, and power supply specifications of the industry. Featuring a standardized pin layout and stackable design, this module achieves cross-vendor compatibility and offers highly flexible, low-risk hardware development solutions for various fields such as smart video analysis devices, industrial AI PCs, embedded edge systems, smart manufacturing, and smart education.
ESWIN Computing’s System-on-Module (SOM) is an embedded core module which integrates edge computing SoCs, memory, storage, power management, PCIe, USB, Ethernet, and other peripheral interfaces into a single, compact design. Combining standardized interfaces and the custom carrier board, this module provides flexible and expandable embedded solutions that allow rapid deployment, making it suitable for a range of fields such as smart video analysis, embedded edge systems, smart manufacturing, and smart education.
As a hub of smart education, the centralized education management cloud platform provides core capabilities such as centralized high-reliability storage, real-time monitoring and smart scheduling, and remote Ops. Smart edge devices are integrated to form a cloud-edge-terminal collaboration system, which lays a core foundation for smart education solutions and provides secure and reliable technical support for the digital transformation of education.
High Computing Power at the Edge Supports various data types including FP16, INT16, and INT8 precision, and offers processing power of 20TOPS at INT8 precision, meeting the edge deployment needs of complex AI models.
Multitask Parallel Processing Supports video decoding up to 32 channels of 1080p@30fps and AI reasoning tasks, and enables response in milliseconds in teaching settings.
Multimodal Algorithm Fusion Leverages AI technologies such as NLP, computer vision, and speech recognition to accurately capture the behaviors and characteristics of teachers and students and conduct real-time analysis of classroom behaviors.
Distributed Cluster Deployment Supports cluster deployment of multiple edge computing devices and implements unified Ops management through the cloud platform.
Cloud-Terminal Collaborative Ops Provides Ops management support such as remote algorithm deployment and firmware update to ensure stable and efficient operations of the system.
By leveraging cloud collaboration technology, this system keeps the main classroom and interactive/audience classroom in sync under various network environments, ensuring stable and reliable connections for cross-regional, real-time interaction between teachers and students. It overcomes the traditional constraints of time and space in education and facilitates the implementation of teaching, teaching-research collaboration, and teaching management. This significantly expands the reach of high-quality educational resources and pushes for substantial progress towards educational equity.
The smart experiment platform leverages deep learning algorithms to perform multimodal data collection and analysis of elements such as the use of experimental equipment, procedural standards, and key operational points. It achieves real-time dynamic tracking and evaluation of physics, chemistry, and biology experiments in middle and high schools. The system can record full-process data from the experiments and accurately keep track of each step. It also offers an objective and fair evaluation system based on AI scoring algorithms, which supports formative assessment and standardized testing in experiment teaching.
This system uses non-contact data acquisition technology to perform multi-dimensional analysis of the entire classroom teaching process, encompassing a total of 12 modules, namely, teaching behavior analysis, teacher movement tracking, classroom language feature recognition, time allocation optimization, speech rate and rhythm assessment, student-teacher behavior pattern evaluation, Q&A interaction analysis, classroom engagement monitoring, quality assessment of teacher-student interactions, attention allocation, student-centered teaching effectiveness evaluation, and student behavior feature analysis. The system provides comprehensive data support for improvement of teaching quality, evidence-based educational research and innovation, and informed decision-making in teaching management, forming a closed smart education loop of data collection, data analysis, and quality optimization.
京公网安备 11030102011303号