1. Introduction
With the development of artificial intelligence and people’s increasing emphasis on education, smart education has been drawing more attention in recent decades [
1]. In recent years, various novel teaching manners have been utilized in college classrooms that leverage multimedia techniques, including textbooks, courseware, video, and voice, among other forms, rather than traditional methods such as blackboard writing. In these innovated educational methodologies, text is no longer the main form of knowledge dissemination, and multi-modal data such as pictures and audio are more conducive to students’ understanding of knowledge [
2,
3,
4]. Therefore, we need more intelligent methods/systems, through which to store, manage, and apply these multi-modal data.
Knowledge graphs serve as an important method, through which data can be organized and managed that interlinks heterogeneous data from different domains [
5]. In the education field, knowledge graphs are often used for teaching and learning in schools. However, these knowledge graphs are frequently constructed manually, consuming a lot of resources, and they cannot be extended to other entities and relationships. Researchers have begun to focus on the automatic construction of educational knowledge graphs. Recent research [
6,
7,
8] used knowledge graphs for ontology construction and achieved some success. Liu et al. predicted the potential relationship between the concept and the course by mapping an online course to the general space of the concept [
9]. Chen et al. proposed a system to construct educational knowledge graphs for students [
10].
In general, most of the previous research data come from online education resources that are not integrated with real classrooms. Traditional educational knowledge graphs only utilize text as the only organizational form, which is monotonous and incomplete for the presentation of concepts or entity information. Compared to text, the use of pictures, the teacher’s voice, and other modes of information make it easier for students to be interested in and understand the information being given to them in class. Therefore, the construction of multi-modal education knowledge graphs is particularly necessary and meaningful.
To tackle the challenges above, we propose a method that is able to automatically construct knowledge graphs that integrate multi-modal teaching resources, such as teacher speech. Taking a data structure course as an example, we used our method to realize the automatic integration of multi-modal educational resources. To improve the professionalism and domain of the educational knowledge graph, we propose a new model for educational entity recognition called EduBERT-BiLSTM-CRF. First, we build an educational lexicon and feed this into the fine-tuned BERT, which allows the BERT model to adaptively learn specific knowledge from the education field. Then, we use BiLSTM to extract the contextual features of each word in the input sentences. A CRF layer is added to obtain the optimal prediction sequence needed to complete education concept recognition. In addition, we use the location information of the entity to construct more accurate semantic relationships for these educational concepts. Finally, as entities in the graph occur in speech data, we convert classroom speech into text through speech recognition technology and link it to the knowledge graph as an entity. We also conduct extensive experiments, and the results show the effectiveness of our method. In summary, the main contributions of this research are as follows:
We propose a model to automatically construct a multi-modal educational knowledge graph, and we provide a way for speech fusion to incorporate and refine the knowledge graph by treating speech as an entity;
We propose a lexicon-based BERT model for educational concept recognition by combining the BiLSTM-CRF model that can better identify educational concepts. For relation extraction, in order to better combine the domain information, we combine the location information of the entity with BERT to dig out the implicit relationships between these entities;
We take computer courses as an example to verify the scalability and feasibility of our work. In addition, the empirical results show that our proposed approach performs competitively better than the state-of-the-art models in entity recognition and in relation extraction.
The rest of this paper is organized as follows:
Section 2 introduces related work on knowledge graph.
Section 3 briefly shows the details, which describes how the multi-nodal knowledge graph was built.
Section 4 presents the experimental results. We summarize this research and discuss the prospects for future research in
Section 5.
Author Contributions
N.L.: Conceptualization, Methodology, Software, Data curation, Writing—Original draft preparation, Writing—Reviewing and Editing; Q.S.: Visualization, Writing—Reviewing and Editing; R.S.: Software, Validation; Y.C.: Investigation, Software; H.X.: Supervision. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the National Natural Science Foundation of China (62077027), the Ministry of Science and Technology of the People’s Republic of China(2018YFC2002500), the Jilin Province Development and Reform Commission, China (2019C053-1), the Education Department of Jilin Province, China (JJKH20200993K), the Department of Science and Technology of Jilin Province, China (20200801002GH), and the European Union’s Horizon 2020 FET Proactive project “WeNet-The Internet of us” (No. 823783).
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The data underlying this article are available in the article.
Acknowledgments
The authors would like to thank all of anonymous reviewers and editors for their helpful suggestions for the improvement of this paper.
Conflicts of Interest
The authors declare no conflict of interest.
References
- Martín, A.C.; Alario-Hoyos, C.; Kloos, C.D. Smart Education: A Review and Future Research Directions. Proceedings 2019, 31, 57. [Google Scholar] [CrossRef] [Green Version]
- D’Mello, S.K.; Olney, A.M.; Blanchard, N.; Samei, B.; Sun, X.; Ward, B.; Kelly, S. Multimodal Capture of Teacher-Student Interactions for Automated Dialogic Analysis in Live Classrooms. In ICMI’15, Proceedings of the 2015 ACM on International Conference on Multimodal Interaction; Association for Computing Machinery: New York, NY, USA, 2015; pp. 557–566. [Google Scholar]
- Suresh, A.; Sumner, T.; Jacobs, J.; Foland, B.; Ward, W. Automating Analysis and Feedback to Improve Mathematics Teachers’ Classroom Discourse. In Proceedings of the AAAI Conference on Artificial Intelligence, Honolulu, HI, USA, 27 January–1 February 2019; 2019; Volume 33, pp. 9721–9728. [Google Scholar]
- Anand, R.; Ottmar, E.; Crouch, J.L.; Whitehill, J. Toward Automated Classroom Observation: Predicting Positive and Negative Climate. In Proceedings of the 2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019), Lille, France, 14–18 May 2019; pp. 1–8. [Google Scholar]
- Heiko, P. Towards Profiling Knowledge Graphs. In Proceedings of the PROFILES@ISWC, Vienna, Austria, 22 October 2017. [Google Scholar]
- Wang, S.; Liang, C.; Wu, Z.; Williams, K.; Pursel, B.; Brautigam, B.; Saul, S.; Williams, H.; Bowen, K.; Giles, C.L. Concept Hierarchy Extraction from Textbooks. In Proceedings of the DocEng’15—ACM Symposium on Document Engineering 2015, Lusanne, Switzerland, 8–11 September 2015. [Google Scholar]
- Lu, W.; Zhou, Y.; Yu, J.; Jia, C. Concept Extraction and Prerequisite Relation Learning from Educational Data. Proc. Conf. AAAI Artif. Intell. 2019, 33, 9678–9685. [Google Scholar] [CrossRef]
- Hu, J.; Zheng, L.; Xu, B. An Approach of Ontology Based Knowledge Base Construction for Chinese K12 Education. In Proceedings of the 2016 First International Conference on Multimedia and Image Processing (ICMIP), Bandar Seri Begawan, Brunei, 1–3 June 2016; pp. 83–88. [Google Scholar]
- Liu, H.; Ma, W.; Yang, Y.; Carbonell, J.G. Learning Concept Graphs from Online Educational Data. J. Artif. Intell. Res. 2016, 55, 1059–1090. [Google Scholar] [CrossRef] [Green Version]
- Chen, P.; Yu, L.; Zheng, V.W.; Chen, X.; Li, X. An Automatic Knowledge Graph Construction System for K-12 Education. In Proceedings of the Fifth Annual ACM Conference on Learning at Scale 2018, London, UK, 26–28 June 2018. [Google Scholar]
- Biega, J.A.; Kuzey, E.; Suchanek, F.M. Inside YAGO2s: A Transparent Information Extraction Architecture. In Proceedings of the 22nd International Conference on World Wide Web—WWW’13 Companion, Rio de Janeiro, Brazil, 13–17 May 2013; pp. 325–328. [Google Scholar]
- Bizer, C.; Lehmann, J.; Kobilarov, G.; Auer, S.; Becker, C.; Cyganiak, R.; Hellmann, S. DBpedia—A crystallization point for the Web of Data. J. Web Semant. 2009, 7, 154–165. [Google Scholar] [CrossRef]
- Fredo, E.; Günther, M.; Krötzsch, M.; Mendez, J.; Vrandečić, D. Introducing Wikidata to the Linked Data Web; Springer: Berlin/Heidelberg, Germany, 2014. [Google Scholar]
- Liu, S.; Yang, H.; Li, J.; Kolmanič, S. Preliminary Study on the Knowledge Graph Construction of Chinese Ancient History and Culture. Information 2020, 11, 186. [Google Scholar] [CrossRef] [Green Version]
- Su, Y.; Zhang, Y. Automatic Construction of Subject Knowledge Graph based on Educational Big Data. In Proceedings of the ICBDE’20—3rd International Conference on Big Data and Education, London, UK, 1–3 April 2020. [Google Scholar]
- Yang, Y.; Liu, H.; Carbonell, J.G.; Ma, W. Concept Graph Learning from Educational Data. In Proceedings of the Eighth ACM International Conference on Web Search and Data Mining, Shanghai, China, 2–6 February 2015. [Google Scholar]
- Senthilkumar, R.D. Concept Maps in Teaching Physics Concepts Applied to Engineering Education: An Explorative Study at The Middle East College, Sultanate of Oman. In Proceedings of the 2017 IEEE Global Engineering Education Conference, EDUCON 2017, Athens, Greece, 25–28 April 2017; pp. 107–110. [Google Scholar]
- Chen, L.; Ye, L.; Wu, Z.; Pursel, B.; Lee, C.G. Recovering Concept Prerequisite Relations from University Course Dependencies. Proc. AAAI Conf. Artif. Intell. 2017, 31, 10550. [Google Scholar]
- Kai, S.; Liu, Y.; Guo, Z.; Wang, C. Visualization for Knowledge Graph Based on Education Data. Int. J. Softw. Inform. 2017, 10, 3. [Google Scholar]
- Zheng, Y.; Liu, R.; Hou, J. The Construction of High Educational Knowledge Graph Based on MOOC. In Proceedings of the 2017 IEEE 2nd Information Technology, Networking, Electronic and Automation Control Conference, ITNEC, Chengdu, China, 15–17 December 2017; pp. 260–263. [Google Scholar]
- Dang, F.; Tang, J.-T.; Pang, K.; Wang, T.; Li, S.; Li, X. Constructing an Educational Knowledge Graph with Concepts Linked to Wikipedia. J. Comput. Sci. Technol. 2021, 36, 1200–1211. [Google Scholar] [CrossRef]
- Yao, S.; Wang, R.; Sun, S.; Bu, D.; Liu, J. Joint Embedding Learning of Educational Knowledge Graphs. arXiv 2019, arXiv:1911.08776. [Google Scholar]
- Lample, G.; Ballesteros, M.; Subramanian, S.; Kawakami, K.; Dyer, C. Neural Architectures for Named Entity Recognition. NAACL. arXiv 2016, arXiv:1603.01360. [Google Scholar]
- Yadav, V.; Bethard, S. A Survey on Recent Advances in Named Entity Recognition from Deep Learning models. arXiv 2018, arXiv:1910.11470. [Google Scholar]
- Nadeau, D.; Sekine, S. A Survey of Named Entity Recognition and Classification. Lingvist. Investig. 2007, 30, 3–26. [Google Scholar] [CrossRef]
- Fu, G.; Kit, C.; Webster, J.J. A Morpheme-based Lexical Chunking System for Chinese. Int. Conf. Mach. Learn. Cybern. 2008, 5, 2455–2460. [Google Scholar]
- Wang, Y.; Wang, L.; Rastegar-Mojarad, M.; Moon, S.; Shen, F.; Afzal, N.; Liu, S.; Zeng, Y.; Mehrabi, S.; Sohn, S.; et al. Clinical information extraction applications: A literature review. J. Biomed. Inform. 2018, 77, 34–49. [Google Scholar] [CrossRef]
- Sutton, C.; McCallum, A. An Introduction to Conditional Random Fields. Found. Trends Mach. Learn. 2012, 4, 267–373. [Google Scholar] [CrossRef]
- Collobert, R.; Weston, J.; Bottou, L.; Karlen, M.; Kavukcuoglu, K.; Kuksa, P.P. Natural Language Processing (Almost) from Scratch. J. Mach. Learn. Res. 2011, 12, 2493–2537. [Google Scholar]
- Zhiheng, H.; Xu, W.; Yu, K. Bidirectional LSTM-CRF Models for Sequence Tagging. arXiv 2015, arXiv:1508.01991. [Google Scholar]
- Yann, D.; Fan, A.; Auli, M.; Grangier, D. Language Modeling with Gated Convolutional Networks. Proc. Mach. Learn. Res. 2017, 70, 933–941. [Google Scholar]
- Hang, Y.; Deng, B.; Li, X.; Qiu, X. TENER: Adapting Transformer Encoder for Named Entity Recognition. arXiv 2019, arXiv:1911.04474. [Google Scholar]
- Hui, C.; Lin, Z.; Ding, G.; Lou, J.-G.; Zhang, Y.; Börje, F.K. GRN: Gated Relation Network to Enhance Convolutional Neural Network for Named Entity Recognition. arXiv 2019, arXiv:1907.05611. [Google Scholar]
- Zhu, Y.; Wang, G.; Börje, F.K. CAN-NER: Convolutional Attention Network for Chinese Named Entity Recognition. arXiv 2019, arXiv:1904.02141. [Google Scholar]
- Peters, M.E.; Neumann, M.; Iyyer, M.; Gardner, M.; Clark, C.; Lee, K.; Zettlemoyer, L. Deep Contextualized Word Representations. NAACL. arXiv 2018, arXiv:1802.05365. [Google Scholar]
- Radford, A.; Karthik, N. Improving Language Understanding by Generative Pre-Training; OpenAI: San Francisco, CA, USA, 2018. [Google Scholar]
- Devlin, J.; Chang, M.-W.; Lee, K.; Toutanova, K. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. arXiv 2019, arXiv:1810.04805. [Google Scholar]
- Aydar, M.; Ozge, B.; Özbay, F. Neural Relation Extraction: A Survey. arXiv 2020, arXiv:2007.04247. [Google Scholar]
- Vu, N.T.; Adel, H.; Gupta, P.; Schütze, H. Combining Recurrent and Convolutional Neural Networks for Relation Classification. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies; Association for Computational Linguistics: San Diego, CA, USA, 2016. [Google Scholar] [CrossRef] [Green Version]
- He, H.; Ganjam, K.; Jain, N.; Lundin, J.; White, R.; Lin, J. An Insight Extraction System on BioMedical Literature with Deep Neural Networks. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing; Association for Computational Linguistics: Copenhagen, Denmark, 2017. [Google Scholar] [CrossRef]
- Zeng, D.; Liu, K.; Chen, Y.; Zhao, J. Distant Supervision for Relation Extraction via Piecewise Convolutional Neural Networks; EMNLP: Lisbon, Portugal, 2015. [Google Scholar]
- Zhang, S.; Zheng, D.; Hu, X.; Yang, M. Bidirectional Long Short-Term Memory Networks for Relation Classification; Fujitsu Research and Development Center: Beijing, China, 2015. [Google Scholar]
- Scarselli, F.; Gori, M.; Tsoi, A.C.; Hagenbuchner, M.; Monfardini, G. The Graph Neural Network Model. IEEE Trans. Neural Netw. 2009, 20, 61–80. [Google Scholar] [CrossRef] [Green Version]
- Zhang, Y.; Qi, P.; Manning, C.D. Graph Convolution over Pruned Dependency Trees Improves Relation Extraction. arXiv 2018, arXiv:1809.10185. [Google Scholar]
- Guo, Z.; Zhang, Y.; Lu, W. Attention Guided Graph Convolutional Networks for Relation Extraction. arXiv 2019, arXiv:abs/1906.07510. [Google Scholar]
- Fu, T.-J.; Li, P.-H.; Ma, W.-Y. GraphRel: Modeling Text as Relational Graphs for Joint Entity and Relation Extraction; Association for Computational Linguistics: Florence, Italy, 2019. [Google Scholar]
- Wu, S.; He, Y. Enriching Pre-trained Language Model with Entity Information for Relation Classification. arXiv 2019, arXiv:1905.08284. [Google Scholar]
- Li, C.; Ye, T. Downstream Model Design of Pre-trained Language Model for Relation Extraction Task. arXiv 2020, arXiv:2004.03786. [Google Scholar]
- Mikolov, T.; Chen, K.; Corrado, G.S.; Dean, J. Efficient Estimation of Word Representations in Vector Space. arXiv 2013, arXiv:1301.3781. [Google Scholar]
Figure 1.
Construction framework of our multi-modal educational knowledge graph.
Figure 2.
The architecture of the EduBERT-BiLSTM-CRF model for educational concept recognition. The input sentences are able to obtain four types of tag sets according to the information in the vocabulary. These vectors are used as the fine-tuned BERT input, and they are then encoded and decoded by BiLSTM and CRF to complete sequence annotation. The Chinese input is “排序是算法” (sorting is an algorithm).
Figure 3.
An example to illustrate tag collection. The character “泡” occurs in two words, and it occurs in the middle of “冒泡排序” and at the end of “冒泡”. Therefore, the I-label set is “冒泡排序” and the E-label set is “冒泡”. The Chinese input is “冒泡排序是一种算法” (bubble sort is an algorithm).
Figure 4.
The main process of speech recognition.
Figure 5.
(a) The loss value changes with the epoch; (b) model results with learning rate and dropout value of 0.0001 and 0.5, respectively.
Figure 6.
Course knowledge concept query module. Different colors represent different entity type. The arrows between entities represent the relationship between them. We use the “双向链表” (double-linked list) as an example. A portion of the education knowledge graph is shown on the left, and entity properties are displayed on the right.
Figure 7.
Multi-modal concept display module. When learners query breadth first traversal, they can not only see the text entities related to it, but they can also see their voice modal knowledge, that is, their teacher’s voice.
Table 1.
An illustration of the current mask mechanism. The Chinese input is “排序是一种算法” (sorting is an algorithm). The result of the input sentence after word segmentation is several words: “排序” (sort), ”是” (is), ”一种” (a), and “算法” (algorithm). Using the original MASK mechanism, only part of the words can be masked, such as “排[M]” and “[M]法”. Using the current MASK mechanism, the whole word can be masked. For example, “排序” can be masked by the token [M].
Input | 排序是一种算法 (Sorting Is an Algorithm) |
---|
Word Segmentation | 排序 是 一种 算法 |
Original Mask | 排[M] 是 一种 [M]法 |
Current Mask | [M][M] 是 一种 [M][M] |
Table 2.
Educational relationships and their descriptions.
Relation Type | Relation Definition |
---|
Inclusion Relationship | Knowledge point A contains knowledge point B, and knowledge point B is the refinement of knowledge point A |
Precursor Relationship | Knowledge point A must be learned before learning knowledge point B |
Identity Relationship | Knowledge point A and knowledge point B are different descriptions of the same knowledge |
Sister Relationship | Knowledge point A and knowledge point B have the same parent knowledge point C, and there is no learning sequence |
Correlation Relationship | Knowledge point A and knowledge point B do not conform to the previous relationships, although they are still relevant |
Table 3.
Comparison of experimental results of the BiLSTM-CRF model, BERT-BiLSTM-CRF model and our model. We ran all models 20 times and took the average. The results indicated statistical significance base on Student’s t-test (p < 0.05).
Method | P | R | F1 |
---|
BiLSTM-CRF | 80.04% | 82.07% | 81.06% |
BERT-BiLSTM-CRF | 80.34% | 85.49% | 82.83% |
EduBERT-BiLSTM-CRF | 85.32% | 85.72% | 85.52% |
Table 4.
Comparison of experimental results of our model with different learning rates. We ran all models 20 times and took the average. The results indicated statistical significance based on Student’s t-test (p < 0.05).
Learning Rate | F1 |
---|
0.01 | 82.43% |
0.001 | 83.97% |
0.0001 | 85.61% |
Table 5.
Comparison of experimental results of our model with different dropout values. We ran all models 20 times and took the average. The results indicated statistical significance based on Student’s t-test (p < 0.05).
Dropout | F1 |
---|
0.1 | 84.92% |
0.3 | 85.31% |
0.5 | 85.78% |
0.7 | 85.12% |
0.9 | 84.72% |
Table 6.
Comparison of experimental results of different methods. We ran all models 20 times and took the average. The results indicated statistical significance based on Student’s t-test (p < 0.05).
Method | Acc | P | R | F1 |
---|
BiLSTM | 64.94% | 67.49% | 81.68% | 73.91% |
CNN | 70.92% | 76.25% | 79.81% | 77.99% |
PCNN | 73.80% | 76.33% | 81.45% | 78.80% |
Lexicon + R-BERT | 75.26% | 76.38% | 84.85% | 80.39% |
Table 7.
Comparison of experimental results of Lexicon + R-BERT model with different dropout values. We ran all models 20 times and took the average. The results indicated statistical significance based on Student’s t-test (p < 0.05).
Dropout | F1 |
---|
0.1 | 79.14% |
0.3 | 79.87% |
0.5 | 80.39% |
0.7 | 79.92% |
0.9 | 78.67% |
| Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).