IFLYTEK(002230)2024-08-01 13:33Education Sector - AI learning machines maintained a high growth rate with sales increasing over 100% from January to May 2024 [1] - The user net promoter score (NPS) remained the highest in the industry [1] - In 2023, the company was the first to implement large model technology in learning machines, launching nine AI 1-on-1 tutoring features [1] - The T30 Ultra, T20 series, and S30 models are fully equipped with these large model features [1] - On June 27, 2024, the learning machine upgraded its AI 1-on-1 tutoring function, enabling multi-modal heuristic explanations, personalized answers, interactive exploratory learning, and hyper-realistic guided tutoring [1] AI-Powered Grading and Analysis - On June 27, 2024, the company launched the Spark Intelligent Grading Machine, which integrates intelligent grading, learning analysis, and personalized homework [2] - The grading machine supports free formatting and unlimited paper sizes, and can intelligently grade multiple subjects and question types [2] - It generates multi-dimensional learning reports instantly, saving significant time in the grading process [2] - In pilot applications, grading time reduced from 90 minutes to 5 minutes, and learning analysis time reduced from 60 minutes to 1 minute [2] - The error resolution rate for students increased from 50% to 73% [2] Energy Sector Applications - The Spark large model has been adopted by industry leaders such as China National Petroleum Corporation (CNPC), National Energy Group, and China Coal [2] - In May 2024, CNPC partnered with the company to develop the "Kunlun Large Model," applying large model technology in oil and gas exploration, refining, and new energy sectors [2] - The company collaborated with National Energy Group to build an industry-specific large model for coal, power, transportation, and chemical sectors, driving efficiency improvements [2] - The China Coal project leveraged the Spark model to enhance procurement efficiency and quality through intelligent solutions [2] International Expansion and Research - On July 19, 2024, the company established an international headquarters in Hong Kong through the Cyberport Easy Landing Program [3] - The company plans to advance core technologies such as intelligent speech and cognitive large models, developing vertical applications tailored to Hong Kong's scenarios [3] - It will strengthen large model and computing power infrastructure, including industry-specific large models, enterprise-specific large models, and high-performance heterogeneous computing platforms [3] - The company will establish an international research institute in Cyberport, focusing on AI and healthcare integration, including medical large models and multi-modal algorithms [3] - Future plans include collaborating with high-end AI institutions and universities in Hong Kong and overseas, expanding international business in education and consumer services, and attracting capital from high-value markets in Southeast Asia and Europe [3]