Abstract
Knowledge tracing (KT) aims to model learners’ knowledge level and predict future performance given their past interactions in learning applications. Adaptive learning systems mainly generate course recommendations based on learner’s knowledge level acquired by KT. However, for KT tasks, learners’ forgetting has not been well modeled. In addition, learner’s individual differences also influence the accuracy of knowledge level prediction. While for recommendation tasks, most of methods are conducted separately from KT tasks, ignoring the deep connection between them. In this paper, we are motivated to propose a Knowledge-Enhanced Multi-task Learning model for Course Recomme-ndation (KMCR), which regards the improved knowledge tracing task (IKTT) as an auxiliary task to assist the primary course recommendation task (CRT). Specifically, in IKTT, for assessing dynamic evolving knowledge level, we not only design a personalized controller to enhance the deep knowledge tracing model for modeling learner’s forgetting behavior, but also use personality to model the individual differences based on the theory of cognitive psychology. In CRT, we adaptively combine learner’s knowledge level obtained by IKTT with their sequential behavior to generate learners’ representation. The experimental results on real-world datasets demonstrate that our approach outperforms related methods in terms of recommendation accuracy.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The value of each personality dimension (on a [1–7] scale) is the average of two related problems (i.e., the positive score and 8 minus the corresponding reverse score).
- 2.
13 indicates all the possible value (1,1.5,...,6.5,7), 5 denotes the class of personality (O,C,E,A,N). The embedding \( e_p* \) is the concatenation of personality value embedding and personality class embedding.
- 3.
To clean the data, we first excluded learners whose average learning time shorter than 30 min and total learning days no more than a week based on their behavior and corresponding time context. We then filtered out all of the contradictory records (e.g., a user rated both 1 or 7 on two opposite statements “I think I am extraverted, enthusiastic” and “I think I am reserved, quiet”) by analyzing their answer to questionnaire.
- 4.
M and SD refer to Mean and Standard Deviation respectively.
- 5.
- 6.
The predicted absolute error = |ground_truth - \(predict\_value\)|, and the mean of predicted absolute error = average(predicted absolute error), the standard deviation of predicted absolute error = standard deviation(predicted absolute error). The ground_truth is learner’s knowledge level \(y_{aux}\) (i.e., the accuracy of course), and the \(predict\_value\) is the predicted knowledge level \(\hat{y}_{aux}\) by KMCR-P or KMCR-K.
References
Beutel, A., et al.: Latent cross: making use of context in recurrent recommender systems. In: WSDM, pp. 46–54 (2018)
Carbonell, J.R.: AI in CAI: an artificial-intelligence approach to computer-assisted instruction. IEEE Trans. Human-Mach. Syst. 11(4), 190–202 (1970)
Corbett, A., et al.: Knowledge tracing: modeling the acquisition of procedural knowledge. User Model. User Adapt. Interact. 4(4), 253–278 (1994)
Dewaele, J.M.: Personality in second language acquisition. In: The Encyclopedia of Applied Linguistics, pp. 4382–4389
Duff, A., et al.: The relationship between personality, approach to learning and academic performance. Personal. Individ. Diff. 36(8), 1907–1920 (2004)
Gan, W., et al.: Modeling learner’s dynamic knowledge construction procedure and cognitive item difficulty for knowledge tracing. Appl. Intell. 50(11), 3894–3912 (2020)
Gosling, S.D., et al.: A very brief measure of the big-five personality domains. J. Res. Personal. 37(6), 504–528 (2003)
Hadash, G., et al.: Rank and rate: multi-task learning for recommender systems. In: RecSys, pp. 451–454 (2018)
He, X., et al.: Neural collaborative filtering. In: WWW, pp. 173–182 (2017)
Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural Comput. 9(8), 1735–1780 (1997)
Huo, Y., et al.: Knowledge modeling via contextualized representations for ISTM-based personalized exercise recommendation. Inf. Sci. 523, 266–278 (2020)
Järvelin, K., et al.: IR evaluation methods for retrieving highly relevant documents. In: SIGIR, vol. 51, pp. 243–250. ACM, New York (2017)
John, O.P., Srivastava, S., et al.: The big five trait taxonomy: history, measurement, and theoretical perspectives. Handbook Personal. Theory Res. 2, 102–138 (1999)
Kerz, E., Wiechmann, D., Silkens, T.: Personality traits moderate the relationship between statistical learning ability and second-language learners’ sentence comprehension. In: CogSci (2020)
Mehta, Y., Majumder, N., Gelbukh, A., Cambria, E.: Recent trends in deep learning based personality detection. Artificial Intelligence Review 53(4), 2313–2339 (2019). https://doi.org/10.1007/s10462-019-09770-z
Meng, W., et al.: Incorporating user micro-behaviors and item knowledge into multi-task learning for session-based recommendation. In: SIGIR, pp. 1091–1100 (2020)
Nagatani, K., et al.: Augmenting knowledge tracing by considering forgetting behavior. In: WWW, pp. 3101–3107 (2019)
Pandey, S., Srivastava, J.: RKT: Relation-aware self-attention for knowledge tracing. In: CIKM, pp. 1205–1214 (2020)
Piech, C., et al.: Deep knowledge tracing. Adv. Neural Inf. Process. Syst. 28, 505–513 (2015)
Rendle, S.: Factorization machines with libfm. TIST 3(3), 1–22 (2012)
Rendle, S., et al.: BPR: Bayesian personalized ranking from implicit feedback. arXiv preprint arXiv:1205.2618 (2012)
Tay, Y., et al.: Multi-pointer co-attention networks for recommendation. In: SIGKDD, pp. 2309–2318 (2018)
Thaker, K., et al.: Recommending remedial readings using student knowledge state. In: EDM, pp. 233–244. ERIC (2020)
Wu, W., et al.: Inferring students’ personality from their communication behavior in web-based learning systems. Int. J. Artif. Intell. Educ. 29(2), 189–216 (2019)
Wu, Z., et al.: Exercise recommendation based on knowledge concept prediction. Knowl. Based Syst. 210, 106481 (2020)
Yu, Z., et al.: Adaptive user modeling with long and short-term preferences for personalized recommendation. In: IJCAI, pp. 4213–4219 (2019)
Zhang, Y., Yang, Q.: A survey on multi-task learning. IEEE Trans. Knowl. Data Eng. (2021)
Zhou, G., et al.: Deep interest evolution network for click-through rate prediction. In: AAAI, vol. 33, pp. 5941–5948 (2019)
Acknowledgements
We thank editors and reviewers for their suggestions and comments. We also thank National Natural Science Foundation of China (under project No. 61907016) and Science and Technology Commission of Shanghai Municipality (under projects No. 21511100302, No. 19511120200 and 20dz2260300) for sponsoring the research work.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Ban, Q., Wu, W., Hu, W., Lin, H., Zheng, W., He, L. (2022). Knowledge-Enhanced Multi-task Learning for Course Recommendation. In: Bhattacharya, A., et al. Database Systems for Advanced Applications. DASFAA 2022. Lecture Notes in Computer Science, vol 13246. Springer, Cham. https://doi.org/10.1007/978-3-031-00126-0_6
Download citation
DOI: https://doi.org/10.1007/978-3-031-00126-0_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-00125-3
Online ISBN: 978-3-031-00126-0
eBook Packages: Computer ScienceComputer Science (R0)