Overview of This Article
4000
Words
Introduction to Knowledge Graphs
A knowledge graph is a large-scale semantic network that provides a semantic representation of the real world. It represents entities as nodes and the attributes of those entities and the relationships between them as edges, forming a web-like graph structure. This structured form is recognizable to humans and friendly to machines, facilitating machine understanding. The large-scale concepts, attributes, and relationships between entities in the graph provide rich semantic and relational information, inherently possessing various characteristics of graphs that can be utilized for related operations and applications. Once constructed, it can also serve as background knowledge for downstream applications. The concept of the knowledge graph was initially proposed by Google, primarily to enhance search engine capabilities and improve search quality. Due to its unique characteristics, the knowledge graph plays an important role in various aspects of artificial intelligence.
Building Subject Knowledge Graphs in Education
The construction of knowledge graphs is typically application-oriented. Before constructing the graph, it is essential to clarify the problems to be solved and the knowledge required to address these problems, and then design the ontology of the knowledge graph to cover this knowledge. First, let’s introduce some common application scenarios and logic of educational knowledge graphs, and based on these applications, design the ontology of the educational graph, followed by the further construction of the knowledge graph.
(1) Application Logic of Educational Knowledge Graphs
The educational knowledge graph centers around subject knowledge, establishing hierarchical relationships between various knowledge points, the associations between knowledge points, and the sequential relationships among different knowledge points, forming a subject knowledge graph. Using this graph, the relationships between knowledge points can be visually presented to students, making it clear and intuitive, naturally helping students build knowledge systems, reference key knowledge points, discover associations between knowledge points, and assist students in summarizing and solidifying their learning, thereby eliminating knowledge gaps.
Once the subject knowledge graph is constructed, it can be associated with teaching resources (textbooks, exam questions, lecture notes, teaching videos, exam papers, etc.), and through user information and learning records, establish connections between knowledge points and users. The knowledge graph can more accurately depict students’ knowledge mastery and resources. This enables precise assessment of user learning conditions, learning path planning, and personalized recommendations for learning resources.
It can also help teachers better understand students’ learning situations, optimize teaching methods, and adjust teaching strategies. By associating with teaching research materials, it can proactively recommend teaching research resources to enhance the efficiency and quality of teachers’ preparation. A teaching assistance Q&A system based on knowledge graph question-answering technology can effectively reduce the burden of simple repetitive questions on teachers and significantly meet students’ inquiry needs.
(2) Educational Knowledge Resource Construction Centered on Subject Knowledge Graphs
Using knowledge graphs as the core, educational knowledge resource construction establishes associations between domain knowledge, linking knowledge points with different versions of textbooks, teaching aids, lecture notes, videos, exam questions, and various educational resources, forming a comprehensive network. These associative networks support upper-layer applications.
In the subject knowledge graphs of the education field, the relationships between knowledge mainly include hierarchical relationships, primarily between parent and child concepts, and between concepts and entities. The concept graph expresses the relationships between concepts in the education field. There are many conceptual contents in the education field, and these relationships form the entire knowledge framework. Inclusion relationships refer to specific sub-concepts under a knowledge point, representing the whole-part relationship. Sequential relationships can be used for learning planning. In different subjects, there are also some special relationships, such as mutual exclusion and causation, which need to be refined and clarified by domain experts and knowledge engineers during the actual construction of the graph resources.
Knowledge in educational graphs also has rich attributes, such as common attributes like “exam points,” “difficult points,” “common mistakes,” and “exam requirements.” Different subjects have specific fine-grained attributes, such as “definitions,” “properties,” “area formulas,” and “perimeter formulas” in mathematics.
Applications of Knowledge Graphs in Smart Education
(1) Precision Profiling Based on Educational Knowledge Graphs
Precision profiling based on educational graphs extracts key information by analyzing user behavior information and its connections with various resource objects to characterize users. User profiles based on knowledge graphs can enhance user profile data, providing a more comprehensive and accurate depiction of users.
1. Accurate User Profiles
User profiling technology is likely familiar to many, as it aims to better characterize users for user understanding, essentially “tagging” users. The accuracy of user profiling determines whether the understanding of users is accurate. In practical applications, two main issues need to be addressed: the first is incomplete profile data, and the second is incorrect profile data.
To address these issues, knowledge graphs can be utilized. The nodes in the knowledge graph abstract educational domain knowledge, covering a sufficient number of entities and concepts, which can serve as a source for user profile tags. The high quality ensures that the tags applied are more accurate. These tags are interconnected, and the graph contains rich semantic relationships, aiding machines in understanding the meaning of these tags. The user-friendly structure facilitates better human understanding and the intuitive discovery of relationships between tags. Algorithms such as tag propagation and cross-domain recommendation can be used to mine more precise tags to describe users, enriching user tags and enhancing user accuracy. Additionally, user profiles can be visually presented as a product, utilizing the relationships within the graph to generate personalized, dynamically changing user knowledge graphs.
2. Accurate Learning Situation Analysis
Knowledge graphs can facilitate more precise learning situation analysis. Traditional educational experts (teachers’ experience) rely heavily on experience to assess learners’ knowledge and ability status, lacking integration of educational measurement theories, which introduces significant randomness and subjectivity. Based on knowledge graphs and big data analysis, methods can objectively mine learners’ learning processes from multiple dimensions. The data can be analyzed from various perspectives, not limited to test scores, error logs, and learning records, but also uncovering deep-seated learning speeds, learning preferences, cognitive levels, and other latent features, making the analysis results more personalized and objective.
For unmet learning objectives, knowledge graphs can perform root cause analysis, identifying weak points and associated knowledge points, effectively addressing gaps. The diagnostic process exhibits better adaptability and personalization.
(2) Enhancing Teaching Quality and Efficiency
Knowledge graphs in auxiliary teaching applications help teachers complete lesson preparation, teaching research, question creation, and exam analysis. The system can recommend similar related materials (lesson plans, course explanation planning, assignments, etc.) to enhance teachers’ teaching efficiency, and graph-based searches can return the required content more accurately.
1. Intelligent Lesson Preparation
By linking subject knowledge points with textbooks, lecture notes, and exercises, the subject knowledge graph can continuously push suitable lesson preparation resources aligned with teachers’ teaching progress and textbook versions, enhancing preparation efficiency and quality. Besides lesson preparation, the subject knowledge graph can also assist in intelligent question creation and exam analysis as background knowledge to complete related tasks, significantly improving research efficiency.
Through precise analysis of learning situations, the system recommends relevant consolidation exercises, formulates targeted teaching strategies, and enhances teaching specificity, enabling precise teaching.
Data mining and intelligence capabilities are comprehensively utilized before, during, and after classes. Before class, data mining techniques are used to obtain student learning situation data and formulate teaching strategies, making decision-making data-driven. During class, targeted teaching is conducted based on established teaching strategies, explaining knowledge points and facilitating group discussions. The knowledge graph visualizes the internal associations of knowledge, helping students build a deep understanding of knowledge. After class, relevant after-class exercises are recommended based on students’ learning situations, learning abilities, and personalized targeted exercises to reinforce mistakes, thereby enhancing the quality and specificity of classroom teaching. The entire teaching process integrates dynamic data analysis and dynamic learning situation diagnosis, realizing tailored instruction and making teaching decisions data-driven and intelligent.
Deep reading based on knowledge graphs primarily aims to achieve an intelligent and comprehensive understanding of the relationships between knowledge. By employing entity linking technology, electronic publications can be recognized and connected, displaying current knowledge information in the form of knowledge cards. It can also relate to other associated knowledge and recommend related knowledge, helping users connect knowledge. This greatly facilitates users’ comprehensive understanding of knowledge. Deep reading can be applied not only in the education field but also effectively in knowledge management and the publishing industry.
The core technology underlying deep reading is entity linking technology. Currently, our Knowledge Workshop has developed an entity understanding service that enables machines to comprehend entities within text, making entity search and semantic search possible, achieving over 90% accuracy and recall in general domains.
Educational robots have become an important application in the education field. Utilizing an education robot centered around a question-answering system can facilitate course Q&A, knowledge retrieval recommendations, teaching management, and other educational tasks. This not only alleviates teachers’ burdens and pressures but also addresses students’ practical issues. An excellent and comprehensive teaching robot integrates multiple system modules, including task-oriented Q&A, chatbots, knowledge-based Q&A, and search recommendation systems, while possessing multi-turn Q&A capabilities. Knowledge graphs play a crucial role in query understanding and knowledge-guided language generation, being central to KBQA.
Evolution Path of Intelligent Education Centered on Knowledge Graphs
Knowledge graphs increasingly take on the mission of facilitating industry intelligence. Exploring the evolution path of industry intelligence based on knowledge graphs is therefore crucial. After years of practice, this path has become clearer, presenting an iterative development model of knowledge resource construction and knowledge application. The basic principles for implementation are overall planning, application-driven, and construction promoted by use. The iterative development path of knowledge resource construction and knowledge application.
(1) Application-Driven, Promoting Construction by Use
For the intelligent implementation of the education industry, it is essential to follow the general industry’s implementation plan, which is application-oriented. With numerous educational application scenarios, overall planning is necessary, gradually implementing based on current business development needs and data, and selecting suitable application scenarios. Knowledge graph capabilities should be built in specific applications that directly generate business value. It is not advisable to blindly construct graphs without specific business support; unless it is a leading enterprise’s AI Lab not pursuing short-term monetization, it is challenging to gain sufficient internal support. Only by generating concrete business value can there be enough motivation to promote the long-term, steady development of intelligence transformation projects centered on knowledge graphs.
(2) Iterative Development
In the intelligent implementation of knowledge graphs in the educational industry (and similarly in vertical industries), compared to knowledge applications, knowledge acquisition and resource construction are greater bottlenecks. Knowledge resource construction is a long-term endeavor that cannot be achieved overnight; it requires solid advancement in knowledge resource construction and knowledge base accumulation without shortcuts. An iterative spiral development model is adopted, starting from specific points to broader applications.
In each iteration cycle, it is crucial to grasp the principle of moderation, prioritizing application scenarios with better expected outcomes, reasonably controlling the boundaries and volume of knowledge, constructing knowledge resources centered around knowledge graphs, and carrying out corresponding knowledge applications. Feedback from internal and external users should be used to improve related applications and knowledge resource construction. After specific applications begin to show results, gradually expand from limited applications to more application scenarios and build more knowledge resources. This entire process continues iterating until comprehensive intelligence is achieved.
Cognitive intelligence is key to advanced artificial intelligence, and the realization of cognitive intelligence relies on knowledge graphs. Nowadays, there are increasing applications of vertical industry intelligence based on knowledge graphs. Many application scenarios in the education field have shown good results. However, the overall application of cognitive intelligence in education is still in the development stage, with many areas to improve. For example, in graph construction, there is a clear demand for multimodal knowledge graphs in educational knowledge graphs, and the complexity of knowledge in the education field increases the difficulty of granularity segmentation of knowledge points, while the richness of relationships between knowledge points also requires continuous optimization. Overall, the application of knowledge graphs in educational intelligence is significant and promising. With the further maturation of technology, support from national policies, continued investment from capital, and ongoing innovation from commercial companies, educational intelligence is expected to achieve remarkable results in the coming years.
