Aiming to solve core problems of ambiguous attention perception and coarse grained intervention in smart education this study proposes a four dimensional measurement and intervention empirical framework for learning efficacy. A layered perception system is constructed by integrating IR PoseNet for spatial orientation and Sub Eye Tracker for visual allocation along with Freq Physio for implicit cognitive load and GL Emotion for emotional utility to break through the limitations of single modal sensing. At the data governance layer the Efficiency Synchronization Algorithm eliminates spatiotemporal asynchrony caused by heterogeneous sampling frequencies of physiological and visual signals while a dynamic weighting mechanism adaptively adjusts modal contributions via scene adaptation factors to enhance robustness in complex environments. Multi scenario empirical results demonstrate that the system achieves high attention recognition accuracy and fake attention detection rates by leveraging the physiology behavior dual evidence chain which effectively resolves information asymmetry in educational supervision and realizes the shift from empirical supervision to data driven governance.

Chang Liu, Hongxia Xiao, Mingyu Lin, Xiaoju Yao. Four-Dimensional Measurement and Intervention of Learning Efficacy. Academic Journal of Business & Management (2026), Vol. 8, Issue 3: 106-113. https://doi.org/10.25236/AJBM.2026.080312.

[1] Ding, J. "The Theoretical Model, Practice Logic and Educational Value of Multimodal  Collaborative Learning Analysis." Journal of Distance Education (1672-0008), 2 (2023).

[2] Xu, Hui, et al. "Analyzing students' attention by gaze tracking and object detection in classroom teaching." Data Technologies and Applications, 57, 5 (2023): 643-667.

[3] Yang, Y, et al." Research on Deep Learning Monitoring Based on Multimodal Data in Intelligent Learning Environment." e-Education Research, 43, 6 (2022):68-76.

[4] Su, Caizhen, et al. "Using synchronized eye movements to predict attention in online video learning." Education Sciences, 14, 5 (2024): 548.

[5] Yin, Rui, et al. "Research Progress and Prospects of Multimodal Learning Analytics Based on Systematic Literature Review." Modern Distance Education Research, 34, 06(2022):54-63.

[6] Liu, Qingtang, et al. "Developments and Prospects of Empirical Research on Multimodal Learning Analysis." e-Education Research, 43, 01(2022):71-78+85.

[7] Kao, Y., Pan, B., Xu, M., Lyu, J., Zhu, X., Chang, Y., ... & Lei, Z.. “Toward 3d face reconstruction in perspective projection: Estimating 6dof face pose from monocular image.” IEEE Transactions on Image Processing, 32, (2023): 3080-3091.

[8] Wan X, Ju J, Tang J, et al. “MPCTrans: Multi-Perspective Cue-Aware Joint Relationship Representation for 3D Hand Pose Estimation via Swin Transformer.” Sensors, 24, 21 (2024): 7029.

[9] Wang Y, Zheng Y. "Intelligent Learning Intervention: Realistic Dilemma, Original Clarification and Proper Form." Open Education Research, 29, 02(2023):103-111.

[10] Vu, Nhung Kim Thi, Thanh Chi Phan, and Thanh Tu Ngo. "Enhance teaching method of self-study competence in the digital environment." Journal of Education and Teaching (JET), 4, 3 (2023): 294-307.

[11] Liu, Hai, et al. "ARHPE: Asymmetric relation-aware representation learning for head pose estimation in industrial human–computer interaction." IEEE Transactions on Industrial Informatics, 18, 10 (2022): 7107-7117.

[12] Wu, Jade, Jingwen Dou, and Sofia Utoft. "Refining Human-Data Interaction: Advanced Techniques for EEGEyeNet Dataset Precision." International Conference on Human-Computer Interaction. Cham: Springer Nature Switzerland, (2024).