(Shibaura Institute of Technology, Japan)(2) Peeraya Sripian
(Shibaura Institute of Technology, Japan)*corresponding author
AbstractEnergy consumption is a growing concern in deep learning, motivating the Green AI paradigm, where models are evaluated not only based on predictive performance but also on energy efficiency. Most existing evaluations focus on spatial tasks and do not fully capture the computational and temporal costs of video-based workflows. This study presents a systematic energy evaluation of spatio-temporal deep learning for video-based student engagement classification. Using the DAiSEE dataset, we compare EfficientNet variants (B0–B7) as frame-level feature extractors and model temporal dynamics with LSTMs on fixed-length sequences. GPU power consumption is measured during training using nvidia-smi, and energy efficiency is quantified with the Kappa Energy Index (KEI), defined as Cohen's Kappa divided by energy consumption (kWh). The results show a clear trade-off between accuracy and energy: EfficientNet-B7 achieves the highest accuracy (0.62) but incurs the highest energy cost (≈5 kWh), resulting in a low KEI, while EfficientNet-B0 achieves competitive accuracy (0.59) with the highest KEI (2.147) due to its low energy consumption (≈0.38 kWh). EfficientNet-B3 strikes a favorable balance (accuracy ≈ 0.61, KEI = 0.747), outperforming larger models under resource constraints. These findings suggest that deeper models do not always maximize energy efficiency, and the KEI value provides a practical metric to guide the selection of energy-efficient models.
KeywordsStudent engagement; spatio-temporal; deep learning; EfficientNet; Green AI; Energy efficiency
|
DOIhttps://doi.org/10.26555/ijain.v12i3.2402 |
Article metricsAbstract views : 102 | PDF views : 8 |
Cite |
Full Text Download
|
References
[1] V. Bolón-Canedo, L. Morán-Fernández, B. Cancela, and A. Alonso-Betanzos, “A review of green artificial intelligence: Towards a more sustainable future,” Neurocomputing, vol. 599, p. 128096, Sep. 2024, doi: 10.1016/j.neucom.2024.128096.
[2] I. Yoon, J. Mun, and K.-S. Min, “Comparative Study on Energy Consumption of Neural Networks by Scaling of Weight-Memory Energy Versus Computing Energy for Implementing Low-Power Edge Intelligence,” Electronics, vol. 14, no. 13, p. 2718, Jul. 2025, doi: 10.3390/electronics14132718.
[3] A. S. Luccioni And A. Hernandez-Garcia, “Counting Carbon: A Survey of Factors Influencing the Emissions of Machine Learning,” arxiv Artif. Intell., pp. 1–19, 2023, [Online]. Available at: https://arxiv.org/pdf/2302.08476.
[4] S. M. Hasan, T. Islam, M. Saifuzzaman, K. R. Ahmed, C.-H. Huang, and A. R. Shahid, “Carbon Emission Quantification of Machine Learning: A Review,” IEEE Trans. Sustain. Comput., vol. 10, no. 6, pp. 1085–1102, Nov. 2025, doi: 10.1109/TSUSC.2025.3578834.
[5] P. Jiang, C. Sonne, W. Li, F. You, and S. You, “Preventing the Immense Increase in the Life-Cycle Energy and Carbon Footprints of LLM-Powered Intelligent Chatbots,” Engineering, vol. 40, pp. 202–210, Sep. 2024, doi: 10.1016/j.eng.2024.04.002.
[6] E. Strubell, A. Ganesh, and A. McCallum, “Energy and Policy Considerations for Modern Deep Learning Research,” Proc. AAAI Conf. Artif. Intell., vol. 34, no. 09, pp. 13693–13696, Apr. 2020, doi: 10.1609/aaai.v34i09.7123.
[7] S. Dash, “Green AI: Enhancing Sustainability and Energy Efficiency in AI-Integrated Enterprise Systems,” IEEE Access, vol. 13, pp. 21216–21228, 2025, doi: 10.1109/ACCESS.2025.3532838.
[8] Y. Yu et al., “Revisit the environmental impact of artificial intelligence: the overlooked carbon emission source?,” Front. Environ. Sci. Eng., vol. 18, no. 12, p. 158, Dec. 2024, doi: 10.1007/s11783-024-1918-y.
[9] C. He, F. Yue, L. Li, Y. Tang, Q. Wu, and W. Chen, “AI carbon footprint: The non-negligible hidden emission source,” Eco-Environment Heal., vol. 4, no. 4, p. 100197, Dec. 2025, doi: 10.1016/j.eehl.2025.100197.
[10] S.-Y. Wang and N.-Z. Ye, “Invisible footprints, visible insights: machine learning reveals Scope 3 emissions,” Front. Sustain., vol. 6, pp. 1–13, Sep. 2025, doi: 10.3389/frsus.2025.1649150.
[11] A. de Vries, “The growing energy footprint of artificial intelligence,” Joule, vol. 7, no. 10, pp. 2191–2194, Oct. 2023, doi: 10.1016/j.joule.2023.09.004.
[12] M. F. Argerich and M. Patiño-Martínez, “Measuring and Improving the Energy Efficiency of Large Language Models Inference,” IEEE Access, vol. 12, pp. 80194–80207, 2024, doi: 10.1109/ACCESS.2024.3409745.
[13] R. Zhang and A. C. S. Chung, “EfficientQ: An efficient and accurate post-training neural network quantization method for medical image segmentation,” Med. Image Anal., vol. 97, p. 103277, Oct. 2024, doi: 10.1016/j.media.2024.103277.
[14] H. Cheng, M. Zhang, and J. Q. Shi, “A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 46, no. 12, pp. 10558–10578, Dec. 2024, doi: 10.1109/TPAMI.2024.3447085.
[15] Q. Li, Z. Chen, Y. Li, Z. Jiang, and S. Cao, “Efficient network compression via gradient-score aware pruning,” Neurocomputing, vol. 660, p. 131870, Jan. 2026, doi: 10.1016/j.neucom.2025.131870.
[16] X. Zhuang et al., “Multi-objective optimization of reservoir development strategy with hybrid artificial intelligence method,” Expert Syst. Appl., vol. 241, p. 122707, May 2024, doi: 10.1016/j.eswa.2023.122707.
[17] B. Tian, Y. Pang, M. Huzaifa, S. Wang, and S. Adve, “Towards Energy-Efficiency by Navigating the Trilemma of Energy, Latency, and Accuracy,” in 2024 IEEE International Symposium on Mixed and Augmented Reality (ISMAR), Oct. 2024, pp. 913–922, doi: 10.1109/ISMAR62088.2024.00107.
[18] Y. Shang, S. Zhou, D. Zhuang, J. Żywiołek, and H. Dincer, “The impact of artificial intelligence application on enterprise environmental performance: Evidence from microenterprises,” Gondwana Res., vol. 131, pp. 181–195, Jul. 2024, doi: 10.1016/j.gr.2024.02.012.
[19] S. Lu et al., “Spatiotemporal Feature Learning for Daily-Life Cough Detection Using FMCW Radar,” Bioengineering, vol. 12, no. 10, p. 1112, Oct. 2025, doi: 10.3390/bioengineering12101112.
[20] M. Wang, J. Xing, J. Su, J. Chen, and L. Yong, “Learning SpatioTemporal and Motion Features in a Unified 2D Network for Action Recognition,” IEEE Trans. Pattern Anal. Mach. Intell., pp. 1–1, 2022, doi: 10.1109/TPAMI.2022.3173658.
[21] Y. Ma, H. Lou, M. Yan, F. Sun, and G. Li, “Spatio-temporal fusion graph convolutional network for traffic flow forecasting,” Inf. Fusion, vol. 104, p. 102196, Apr. 2024, doi: 10.1016/j.inffus.2023.102196.
[22] W. Chen et al., “Spatio-temporal characteristics and influencing factors of traditional villages in the Yangtze River Basin: a Geodetector model,” Herit. Sci., vol. 11, no. 1, p. 111, May 2023, doi: 10.1186/s40494-023-00948-x.
[23] B. Olczak, M. Wilkosz-mamcarczyk, B. Prus, K. Hodor, and R. Dixon-gough, “Application of the building cohesion method in spatial planning to shape patterns of the development in a suburban historical landscape of a ‘village within Kraków,’” Land use policy, vol. 114, p. 105997, Mar. 2022, doi: 10.1016/j.landusepol.2022.105997.
[24] R. Das and S. Dev, “Optimizing student engagement detection using facial and behavioral features,” Neural Comput. Appl., vol. 37, no. 23, pp. 19063–19085, Aug. 2025, doi: 10.1007/s00521-025-11317-z.
[25] I. Qarbal, N. Sael, and S. Ouahabi, “Student’s Engagement Detection Based on Computer Vision: A Systematic Literature Review,” IEEE Access, vol. 13, pp. 140519–140545, 2025, doi: 10.1109/ACCESS.2025.3596885.
[26] T. Shan, S. Feng, K. Li, R. Chang, and R. Huang, “Unveiling the effects of artificial intelligence and green technology convergence on carbon emissions: An explainable machine learning-based approach,” J. Environ. Manage., vol. 373, p. 123657, Jan. 2025, doi: 10.1016/j.jenvman.2024.123657.
[27] R. Ranpara, “Energy-efficient green AI architectures for circular economies through multi-layered sustainable resource optimization framework,” Discov. Sustain., vol. 6, no. 1, p. 1031, Oct. 2025, doi: 10.1007/s43621-025-01846-x.
[28] S. Khan et al., “Green AI techniques for reducing energy consumption in AI systems,” Array, vol. 29, p. 100652, Mar. 2026, doi: 10.1016/j.array.2025.100652.
[29] T. Yigitcanlar, R. Mehmood, and J. M. Corchado, “Green Artificial Intelligence: Towards an Efficient, Sustainable and Equitable Technology for Smart Cities and Futures,” Sustainability, vol. 13, no. 16, p. 8952, Aug. 2021, doi: 10.3390/su13168952.
[30] R. Verdecchia, J. Sallou, and L. Cruz, “A systematic review of Green AI,” WIREs Data Min. Knowl. Discov., vol. 13, no. 4, p. e1507, Jul. 2023, doi: 10.1002/widm.1507.
[31] R. Verdecchia, P. Lago, C. Ebert, and C. de Vries, “Green IT and Green Software,” IEEE Softw., vol. 38, no. 6, pp. 7–15, Nov. 2021, doi: 10.1109/MS.2021.3102254.
[32] O. O. Oyewole and J. F. Joseph, “Sustainable AI and Green Computing: Reducing the Environmental Impact of Large-Scale Models with Energy-Efficient Techniques,” Int. J. Sci. Res. Netw. Secur. Commun., vol. 13, no. 3, pp. 19–26, Jun. 2025, doi: 10.26438/ijsrnsc.v13i3.276.
[33] R. Z. Khudhur and M. A. Mohammed, “A Spatio-Temporal Deep Learning Approach for Efficient Deepfake Video Detection,” ARO-THE Sci. J. KOYA Univ., vol. 13, no. 2, pp. 75–82, Aug. 2025, doi: 10.14500/aro.12190.
[34] Z. Zhang et al., “PromptST: Prompt-Enhanced Spatio-Temporal Multi-Attribute Prediction,” in Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, Oct. 2023, pp. 3195–3205, doi: 10.1145/3583780.3615016.
[35] J. Wu, Z. Niu, X. Li, L. Huang, P. S. Nielsen, and X. Liu, “Understanding multi-scale spatiotemporal energy consumption data: A visual analysis approach,” Energy, vol. 263, p. 125939, Jan. 2023, doi: 10.1016/j.energy.2022.125939.
[36] G. Zhou, O. Soufan, J. Ewald, R. E. W. Hancock, N. Basu, and J. Xia, “NetworkAnalyst 3.0: a visual analytics platform for comprehensive gene expression profiling and meta-analysis,” Nucleic Acids Res., vol. 47, no. W1, pp. W234–W241, Jul. 2019, doi: 10.1093/nar/gkz240.
[37] C.-W. Shen, J.-W. Jiang, and H.-P. Hsieh, “STGym: A Modular Benchmark for Spatio-Temporal Networks With a Survey and Case Study on Traffic Forecasting,” IEEE Trans. Big Data, vol. 12, no. 1, pp. 15–33, Feb. 2026, doi: 10.1109/TBDATA.2025.3618482.
[38] Z. Ma, C. Chen, J. Chen, and Y. Jiang, “STC-SORT: A Dynamic Spatio-Temporal Consistency Framework for Multi-Object Tracking in UAV Videos,” Appl. Sci., vol. 16, no. 2, p. 1062, Jan. 2026, doi: 10.3390/app16021062.
[39] T. Fütterer et al., “Artificial intelligence in classroom management: A systematic review on educational purposes, technical implementations, and ethical considerations,” Comput. Educ. Artif. Intell., vol. 9, p. 100483, Dec. 2025, doi: 10.1016/j.caeai.2025.100483.
[40] A. T. S. and R. M. R. Guddeti, “Automatic detection of students’ affective states in classroom environment using hybrid convolutional neural networks,” Educ. Inf. Technol., vol. 25, no. 2, pp. 1387–1415, Mar. 2020, doi: 10.1007/s10639-019-10004-6.
[41] M. Alruwaili and M. Mohamed, “An Integrated Deep Learning Model with EfficientNet and ResNet for Accurate Multi-Class Skin Disease Classification,” Diagnostics, vol. 15, no. 5, p. 551, Feb. 2025, doi: 10.3390/diagnostics15050551.
[42] W. Lu, Y. Yang, R. Song, Y. Chen, T. Wang, and C. Bian, “A Video Dataset for Classroom Group Engagement Recognition,” Sci. Data, vol. 12, no. 1, p. 644, Apr. 2025, doi: 10.1038/s41597-025-04987-w.
[43] A. Gupta, A. D’Cunha, K. Awasthi, and V. Balasubramanian, “DAiSEE: Towards User Engagement Recognition in the Wild,” J. LATEX Cl. FILES, vol. 14, no. 8, pp. 1–14, 2015, [Online]. Available at: https://arxiv.org/abs/1609.01885.
[44] M. Tan and Q. V. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” 36th Int. Conf. Mach. Learn. ICML 2019, vol. 2019-June, pp. 10691–10700, May 2019. [Online]. Available at: https://arxiv.org/abs/1905.11946v5.
[45] S. R. Iyer, “Green AI: Energy-Efficient Training of Large-Scale Models,” Int. J. Res. Appl. Innov., vol. 8, no. 2, pp. 11947–11951, Mar. 2025. [Online]. Available at: https://www.ijrai.org/index.php/ijrai/article/view/85.
[46] R. Su, L. He, and M. Luo, “Leveraging part-and-sensitive attention network and transformer for learner engagement detection,” Alexandria Eng. J., vol. 107, pp. 198–204, Nov. 2024, doi: 10.1016/j.aej.2024.06.074.
[47] R. Das and S. Dev, “Enhancing frame-level student engagement classification through knowledge transfer techniques,” Appl. Intell., vol. 54, no. 2, pp. 2261–2276, Jan. 2024, doi: 10.1007/s10489-023-05256-2.
[48] Haibo He and E. A. Garcia, “Learning from Imbalanced Data,” IEEE Trans. Knowl. Data Eng., vol. 21, no. 9, pp. 1263–1284, Sep. 2009, doi: 10.1109/TKDE.2008.239.
[49] H. Kaur, H. S. Pannu, and A. K. Malhi, “A Systematic Review on Imbalanced Data Challenges in Machine Learning,” ACM Comput. Surv., vol. 52, no. 4, pp. 1–36, Jul. 2020, doi: 10.1145/3343440.
[50] A. G. Howard et al., “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications,” arXiv Comput. Vis. Pattern Recognit., pp. 1–9, Apr. 2017. [Online]. Available at: https://arxiv.org/abs/1704.04861v1.
[51] A. Dosovitskiy et al., “an Image Is Worth 16X16 Words: Transformers for Image Recognition At Scale,” ICLR 2021 - 9th Int. Conf. Learn. Represent., pp. 1–22, 2021, [Online]. Available at: https://arxiv.org/pdf/2010.11929.
[52] S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory,” Neural Comput., vol. 9, no. 8, pp. 1735–1780, Nov. 1997, doi: 10.1162/neco.1997.9.8.1735.
[53] M. Krichen and A. Mihoub, “Long Short-Term Memory Networks: A Comprehensive Survey,” AI, vol. 6, no. 9, p. 215, Sep. 2025, doi: 10.3390/ai6090215.
[54] M. M. Taye, “Understanding of Machine Learning with Deep Learning: Architectures, Workflow, Applications and Future Directions,” Computers, vol. 12, no. 5, p. 91, Apr. 2023, doi: 10.3390/computers12050091.
[55] T. T. Kieu Tran, T. Lee, J. Y. Shin, J. S. Kim, and M. Kamruzzaman, “Deep learning-based maximum temperature forecasting assisted with meta-learning for hyperparameter optimization,” Atmosphere (Basel)., vol. 11, no. 5, pp. 1–21, 2020, doi: 10.3390/ATMOS11050487.
[56] S. Aquino-Brítez, P. García-Sánchez, A. Ortiz, and D. Aquino-Brítez, “Towards an Energy Consumption Index for Deep Learning Models: A Comparative Analysis of Architectures, GPUs, and Measurement Tools,” Sensors, vol. 25, no. 3, p. 846, Jan. 2025, doi: 10.3390/s25030846.
[57] T. Selim, I. Elkabani, and M. A. Abdou, “Students Engagement Level Detection in Online e-Learning Using Hybrid EfficientNetB7 Together With TCN, LSTM, and Bi-LSTM,” IEEE Access, vol. 10, pp. 99573–99583, 2022, doi: 10.1109/ACCESS.2022.3206779.
[58] A. Sánchez-Mompó, I. Mavromatis, P. Li, K. Katsaros, and A. Khan, “Green MLOps to Green GenOps: An Empirical Study of Energy Consumption in Discriminative and Generative AI Operations,” Information, vol. 16, no. 4, p. 281, Mar. 2025, doi: 10.3390/info16040281.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
___________________________________________________________
International Journal of Advances in Intelligent Informatics
ISSN 2442-6571 (print) | 2548-3161 (online)
Organized by UAD and ASCEE Computer Society
Published by Universitas Ahmad Dahlan
W: http://ijain.org
E: info@ijain.org (paper handling issues)
andri.pranolo.id@ieee.org (publication issues)
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0