Langhui Tech is deeply engaged in education AI data services, building 22 high-quality datasets covering K12 basic education, higher education, vocational education, subject knowledge graphs, and education assessment. With a 10M+ question bank, 500K+ lesson plans, and a 5,000+ certified teacher expert network, we provide high-quality, multi-disciplinary, all-grade corpora for education large model training.
From Khanmigo personalized mentors to GPT-4 educational megamodels, from multimodular teaching to intelligent assessment, AI is deeply reshaping educational ecology
Khan Academy's personalized AI tutor serves 10 million students with all-subject tutoring. Built on GPT-4, Khanmigo uses Socratic dialogue to guide student thinking and has been deployed in thousands of schools across the United States.
OpenAI released an education-specific version of GPT-4, supporting multi-subject tutoring, adaptive learning, and intelligent grading. It outperforms 80% of human test takers on benchmarks such as SAT and AP exams, driving rapid development of education large models.
MIT released a multimodal teaching AI system that supports video, image, and text understanding. It analyzes classroom teaching videos, recognizes student expressions and engagement, and provides teachers with real-time teaching feedback, improving classroom efficiency by 35%.
ETS launched an AI automated scoring system covering standardized tests such as SAT/GRE, achieving scoring consistency above 95%. The AI scoring system has graded over 100 million exam papers, significantly shortening scoring cycles and reducing labor costs.
Stanford released an education knowledge graph covering 100,000+ knowledge points across all K12 subjects, building structured graphs of prerequisite dependencies, relationships, and competency levels to support personalized learning path recommendations and knowledge gap diagnosis.
Pearson launched an AI adaptive learning platform that dynamically adjusts learning paths and difficulty based on learning behavior data. Covering 50 million learners worldwide, it improves average learning efficiency by 40% and knowledge mastery by 25%.
Building an AIS data infrastructure based on credible data
Subject textbooks, exam questions, and lesson plans covering all K12 grades and subjects, including systematic corpora for Chinese, Math, English, Physics, Chemistry, Biology, History, Geography, and other core subjects.
Math/Physics/Chemistry/Biology/History/Chinese, 100,000+ knowledge points, building structured graphs of prerequisite dependencies, relationships, and competency levels.
Scoring standards, learning behaviors, competency diagnosis, and learning analytics — multi-dimensional education assessment data supporting intelligent scoring and learning diagnosis AI training.
5,000+ certified teachers covering all subjects, with expert review and quality control. An expert review team composed of frontline teachers, teaching researchers, and master teachers.
Coverage of K12 All-Scientific + College core
Certification teacher double-check + AI-assisted verification
Primary/ Lower secondary/ Upper secondary/ Higher/ Vocational
5000+Accredited teachers participating in the trial Nuclear
22 high-quality data sets covering the educational AI landscape
Full-text corpora of K12 Chinese textbooks from PEP, Sujiao, and Ministry-compiled editions, including structured annotations for texts, vocabulary, reading comprehension, and writing samples.
Math, Physics, Chemistry, and Biology science textbook corpora, with structured annotations for formulas, theorems, examples, and experiments, covering all K12 grades.
Textbook corpora for English, History, Geography, Politics, and other humanities subjects, with structured annotations for vocabulary, grammar, historical sources, and maps.
Covering all grades from primary to high school, including multiple choice, fill-in-the-blank, and free-response questions, with standard answers and detailed explanations.
Past Gaokao (college entrance exam) questions and mock exams from top schools, with deep annotations including tested knowledge points, difficulty coefficients, solution approaches, and knowledge chains.
Question banks for five major subject competitions: Math, Physics, Chemistry, Biology, and Informatics Olympiads, with annotations for questions, answers, solution methods, and knowledge point analysis.
High-quality lesson plans across all subjects, with structured annotations for teaching objectives, key and difficult points, teaching processes, board designs, and homework assignments.
High-quality courseware PPTs with multimodal content including graphics, animations, and videos, categorized and annotated by subject, chapter, and knowledge point.
Classroom recording videos and transcripts from master teachers, with deep annotations for teaching segments, teacher-student interactions, and teaching strategies.
Knowledge point relationship graphs across Math/Physics/Chemistry/Biology/History/Chinese subjects, including prerequisite dependencies, relationships, and competency levels.
Association annotations between questions and knowledge points, with each question mapped to tested knowledge points, competency levels, and cognitive dimensions, supporting intelligent question generation and diagnosis.
Personalized learning path data based on knowledge graphs, including learning sequences, knowledge point mastery, and recommended paths.
Student online learning behavior data, including multi-dimensional behavior logs for answering questions, watching videos, and interactions, supporting learning analytics AI training.
Student subject competency diagnosis data, with annotations for knowledge point mastery, competency levels, and weak areas, supporting personalized diagnosis AI.
Subjective question scoring standards and samples across subjects, with annotations for scoring rubrics, scoring points, and typical answers, supporting intelligent grading AI training.
Credible data set + customized data services + network of certification experts to build the Education AI full-link data infrastructure
22 ready-to-use standardized education datasets in JSON format, available for immediate use. Covering five major categories: K12 textbooks, question bank, lesson plans, knowledge graph, and assessment data — out-of-the-box corpora for education AI training.
Industrial-grade annotation platform, customizing education annotations on demand. Teacher expert review + AI-assisted annotation, supporting customized needs such as textbook annotation, question analysis, lesson plan generation, and knowledge graph construction.
5,000+ certified teacher expert network, covering all subjects. A quality moat composed of frontline teachers, teaching researchers, and master teachers, providing professional review and quality control for education data annotation.
Six typical AI applications based on the Langhui Education Data Set
Based on learning behavior and competency diagnosis data, building adaptive learning paths and recommending personalized learning content, with 40% improvement in learning efficiency.
AI tutor for all-subject tutoring, based on the Khanmigo paradigm, using Socratic dialogue to guide student thinking, covering all K12 grades.
Intelligent grading system with automated scoring for subjective questions, achieving 95%+ scoring consistency, covering essays, free-response questions, and other types.
Learning analytics and competency diagnosis, multi-dimensional assessment of student learning status, providing teachers with precise teaching decision support.
Subject knowledge point association and dependency analysis, 100K+ knowledge points building an education knowledge network, supporting intelligent question generation and diagnosis.
Lesson plan generation, courseware creation, and instructional design — AI assists teachers in lesson preparation, reducing prep time by 60% and improving teaching quality.
Contact Langhui Data Specialist to obtain a complete catalogue of educational data sets, test privileges and customized solutions