Abstract
Neural networks’ capability to model non-linear relationships strongly depends on their activation functions (AFs). This dependency makes the search for AFs with better performance a relevant open research subject. This research work proposes two Trainable Gaussian-based AFs for Multilayer Perceptron neural networks on sensor-based human activity recognition (HAR), namely, the Four-Parameter Activation Gaussian Radial Basis Function (T4GRBF) and the Weighted Gaussian Radial Basis Function (WGRBF). The T4GRBF considers the training of four parameters. Two of these parameters regulate the center and spread of the Gaussian shape separately. Another parameter regulates the center and shape simultaneously, and the last parameter regulates the output range of the AF. The WGRBF adjusts the Gaussian shape with only two trainable parameters, and it is then scaled by the AF input. The proposals are validated through experiments on the Opportunity and UniMiB SHAR benchmark datasets. The results found for the Opportunity dataset in this research work evidence that the sliding window segmentation method has a high impact on the AFs’ performance, and the results in the UniMiB dataset show that the trainable GRBF-based AFs improve the HAR models performance. Furthermore, in both datasets, the Trainable Gaussian-based AFs fit the training data better than the standard AFs, regardless of the specific time window setup used in the experiments.
Similar content being viewed by others
References
Scardapane S, Van Vaerenbergh S, Totaro S, Uncini A (2019) Kafnets: Kernel-based non-parametric activation functions for neural networks. Neural Netw 110:19–32. https://doi.org/10.1016/j.neunet.2018.11.002
Paduano I, Mileto A, Lofrano E (2023) A perspective on ai-based image analysis and utilization technologies in building engineering: recent developments and new directions. Buildings. https://doi.org/10.3390/buildings13051198
Wang C, Song L-h, Yuan Z, Fan J-s (2023) State-of-the-art ai-based computational analysis in civil engineering. J Ind Inf Integr 33:100470. https://doi.org/10.1016/j.jii.2023.100470
Ren Y, Jin C, Fang S, Yang L, Wu Z, Wang Z, Peng R, Gao K (2023) A comprehensive review of key technologies for enhancing the reliability of lithium-ion power batteries. Energies. https://doi.org/10.3390/en16176144
Cui S, Gao Y, Huang Y, Shen L, Zhao Q, Pan Y, Zhuang S (2023) Advances and applications of machine learning and deep learning in environmental ecology and health. Environ Pollut 335:122358. https://doi.org/10.1016/j.envpol.2023.122358
Amin AA, Sajid Iqbal M, Hamza Shahbaz M (2024) Development of intelligent fault-tolerant control systems with machine learning, deep learning, and transfer learning algorithms: A review. Expert Syst Appl 238:121956. https://doi.org/10.1016/j.eswa.2023.121956
Jieyang P, Kimmig A, Dongkun W, Niu Z, Zhi F, Jiahai W, Liu X, Ovtcharova J (2023) A systematic review of data-driven approaches to fault diagnosis and early warning. J Intell Manuf 34(8):3277–3304. https://doi.org/10.1007/s10845-022-02020-0
Yang D, Zhang W, Jiang Y (2023) Mechanical fault diagnosis based on deep transfer learning: a review. Meas Sci Technol 34(11):112001. https://doi.org/10.1088/1361-6501/ace7e6
Guo Y, Zhang J, Sun B, Wang Y (2023) Adversarial deep transfer learning in fault diagnosis: progress, challenges, and future prospects. Sensors. https://doi.org/10.3390/s23167263
Mishra BK, Thakker D, Mazumdar S, Neagu D, Gheorghe M, Simpson S (2020) A novel application of deep learning with image cropping: a smart city use case for flood monitoring. J Reliab Intell Environ 6(1):51–61. https://doi.org/10.1007/s40860-020-00099-x
Schmitt M (2023) Securing the digital world: protecting smart infrastructures and digital industries with artificial intelligence (ai)-enabled malware and intrusion detection. J Ind Inf Integr 36:100520. https://doi.org/10.1016/j.jii.2023.100520
Gil Pereira P, Vogelgesang K, Miodek M, Schmidt A, Herfet T (2023) Deepsharq: hybrid error coding using deep learning. J Reliab Intell Environ 9(3):283–301. https://doi.org/10.1007/s40860-023-00207-7
Bingham G, Miikkulainen R (2022) Discovering parametric activation functions. Neural Netw 148:48–65. https://doi.org/10.1016/j.neunet.2022.01.001
Hu H, Liu A, Guan Q, Qian H, Li X, Chen S, Zhou Q (2022) Adaptively customizing activation functions for various layers. IEEE Trans Neural Netw Learn Syst. https://doi.org/10.1109/TNNLS.2021.3133263
Tavakoli M, Agostinelli F, Baldi P (2021) Splash: Learnable activation functions for improving accuracy and adversarial robustness. Neural Netw 140:1–12. https://doi.org/10.1016/j.neunet.2021.02.023
Wang X, Ren H, Wang A (2022) Smish: A novel activation function for deep learning methods. Electronics. https://doi.org/10.3390/electronics11040540
Shen S-L, Zhang N, Zhou A, Yin Z-Y (2022) Enhancement of neural networks with an alternative activation function tanhlu. Exp Syst Appl. https://doi.org/10.1016/j.eswa.2022.117181
Li J, Ding F, Hayat T (2022) A novel nonlinear optimization method for fitting a noisy gaussian activation function. Int J Adapt Control Signal Process 36:690–707. https://doi.org/10.1002/acs.3367
Kiliçarslan S, Celik M (2022) Kaf + rsigelu: a nonlinear and kernel-based activation function for deep neural networks. Neural Comput Appl. https://doi.org/10.1007/s00521-022-07211-7
Apicella A, Donnarumma F, Isgrò F, Prevete R (2021) A survey on modern trainable activation functions. Neural Netw 138:14–32. https://doi.org/10.1016/j.neunet.2021.01.026
Essa E, Abdelmaksoud IR (2023) Temporal-channel convolution with self-attention network for human activity recognition using wearable sensors. Knowl-Based Syst 278:110867. https://doi.org/10.1016/j.knosys.2023.110867
Raja Sekaran S, Han PY, Yin OS (2023) Smartphone-based human activity recognition using lightweight multiheaded temporal convolutional network. Expert Syst Appl 227:120132. https://doi.org/10.1016/j.eswa.2023.120132
Choudhury NA, Soni B (2023) An adaptive batch size-based-cnn-lstm framework for human activity recognition in uncontrolled environment. IEEE Trans Industr Inf 19(10):10379–10387. https://doi.org/10.1109/TII.2022.3229522
Jiang Q, Zhu L, Shu C, Sekar V (2022) Multilayer perceptron neural network activated by adaptive gaussian radial basis function and its application to predict lid-driven cavity flow. Acta Mech Sin. https://doi.org/10.1007/s10409-021-01144-5
Augusto JC (2022) Contexts and context-awareness revisited from an intelligent environments perspective. Appl Artif Intell. https://doi.org/10.1080/08839514.2021.2008644
Ferrari A, Micucci D, Mobilio M, Napoletano P (2021) Trends in human activity recognition using smartphones. J Reliab Intell Environ 7(3):189–213. https://doi.org/10.1007/s40860-021-00147-0
Feng H, Shen Q, Song R, Shi L, Xu H (2024) Atfa: adversarial time-frequency attention network for sensor-based multimodal human activity recognition. Expert Syst Appl 236:121296. https://doi.org/10.1016/j.eswa.2023.121296
Sarkar A, Hossain SKS, Sarkar R (2023) Human activity recognition from sensor data using spatial attention-aided cnn with genetic algorithm. Neural Comput Appl 35(7):5165–5191. https://doi.org/10.1007/s00521-022-07911-0
Chen C-T, Chang W-D (1996) A feedforward neural network with function shape autotuning. Neural Netw 9(4):627–641. https://doi.org/10.1016/0893-6080(96)00006-8
Ramachandran P, Zoph B, Le QV (2017) Searching for activation functions
Apicella A, Isgrò F, Prevete R (2019) A simple and efficient architecture for trainable activation functions. Neurocomputing 370:1–15. https://doi.org/10.1016/j.neucom.2019.08.065
Hornik K, Stinchcombe M, White H (1989) Multilayer feedforward networks are universal approximators. Neural Netw 2(5):359–366. https://doi.org/10.1016/0893-6080(89)90020-8
Ferrari A, Micucci D, Mobilio M, Napoletano P (2023) Deep learning and model personalization in sensor-based human activity recognition. J Reliab Intell Environ 9(1):27–39. https://doi.org/10.1007/s40860-021-00167-w
Sanchez S, Machacuay J, Quinde M (2023) Federated learning for human activity recognition on the mhealth dataset. In: Rutkowski L, Scherer R, Korytkowski M, Pedrycz W, Tadeusiewicz R, Zurada JM (eds) Artificial intelligence and soft computing. Springer, Cham, pp 215–225
Gao W, Zhang L, Teng Q, He J, Wu H (2021) Danhar: Dual attention network for multimodal human activity recognition using wearable sensors. Appl Soft Comput 111:107728. https://doi.org/10.1016/j.asoc.2021.107728
Cheng X, Zhang L, Tang Y, Liu Y, Wu H, He J (2022) Real-time human activity recognition using conditionally parametrized convolutions on mobile and wearable devices. IEEE Sens J 22(6):5889–5901. https://doi.org/10.1109/JSEN.2022.3149337
Gu F, Chung M-H, Chignell M, Valaee S, Zhou B, Liu X (2021) A survey on deep learning for human activity recognition. ACM Comput Surv 54:8. https://doi.org/10.1145/3472290
Glorot X, Bengio Y (2010) Understanding the difficulty of training deep feedforward neural networks. In: Teh YW, Titterington M (eds.) Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics. Proceedings of Machine Learning Research, vol. 9, pp. 249–256. PMLR, Chia Laguna Resort, Sardinia, Italy
Chavarriaga R, Sagha H, Calatroni A, Digumarti ST, Tröster G, Millán R, Roggen D (2013) The opportunity challenge: a benchmark database for on-body sensor-based activity recognition. Pattern Recogn Lett 34(15):2033–2042. https://doi.org/10.1016/j.patrec.2012.12.014
Micucci D, Mobilio M, Napoletano P (2017) Unimib shar: A dataset for human activity recognition using acceleration data from smartphones. Appl Sci 7:10. https://doi.org/10.3390/app7101101
Ige AO, Mohd Noor MH (2022) A survey on unsupervised learning for wearable sensor-based activity recognition. Appl Soft Comput 127:109363. https://doi.org/10.1016/j.asoc.2022.109363
Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, Kaiser L, Polosukhin I (2023) Attention is all you need
Kingma DP, Ba J (2017) Adam: a method for stochastic optimization
Loshchilov I, Hutter F (2019) Decoupled weight decay regularization
Du KL, Swamy MNS (2019) Fundamentals of machine learning. Springer, London, pp 21–63. https://doi.org/10.1007/978-1-4471-7452-3_2
Raschka S (2018) Model evaluation, model selection, and algorithm selection in machine learning. arXiv preprint arXiv:1811.12808
Ghosal S, Sarkar M, Sarkar R (2022) Nofed-net: Non-linear fuzzy ensemble of deep neural networks for human activity recognition. IEEE Internet Things J. https://doi.org/10.1109/JIOT.2022.3155560
Ma C, Li W, Cao J, Du J, Li Q, Gravina R (2020) Adaptive sliding window based activity recognition for assisted livings. Inf Fusion 53:55–65. https://doi.org/10.1016/j.inffus.2019.06.013
Noor MHM, Salcic Z, Wang KI-K (2017) Adaptive sliding window segmentation for physical activity recognition using a single tri-axial accelerometer. Pervasive Mob Comput 38:41–59. https://doi.org/10.1016/j.pmcj.2016.09.009
Kwapisz JR, Weiss GM, Moore SA (2011) Activity recognition using cell phone accelerometers. ACM SIGKDD Explor Newsl 12(2):74–82
Reiss A, Stricker D (2012) Introducing a new benchmarked dataset for activity monitoring. In: 2012 16th International Symposium on Wearable Computers, pp. 108–109. IEEE
Anguita D, Ghio A, Oneto L, Parra X, Reyes-Ortiz JL et al (2013) A public domain dataset for human activity recognition using smartphones. Esann 3:3
Ferrari A, Micucci D, Mobilio M, Napoletano P (2020) On the personalization of classification models for human activity recognition. IEEE Access 8:32066–32079. https://doi.org/10.1109/ACCESS.2020.2973425
Ferrari Anna NP, Daniela M (2023) Deep learning and model personalization in sensor-based human activity recognition. J Reliable Intell Environ 9:27–39. https://doi.org/10.1007/s40860-021-00167-w
Author information
Authors and Affiliations
Contributions
JM developed the idea of the proposed AFs, coded the experiments to validate the AFs and wrote the main manuscript text. MQ designed the methodology to validate the proposed AFs and support the manuscript’s writing. All authors reviewed the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Machacuay, J., Quinde, M. Trainable Gaussian-based activation functions for sensor-based human activity recognition. J Reliable Intell Environ (2024). https://doi.org/10.1007/s40860-024-00221-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s40860-024-00221-3