{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,15]],"date-time":"2026-08-15T04:52:24Z","timestamp":1786769544473,"version":"3.56.0"},"reference-count":51,"publisher":"Wiley","issue":"1","license":[{"start":{"date-parts":[[2025,4,21]],"date-time":"2025-04-21T00:00:00Z","timestamp":1745193600000},"content-version":"vor","delay-in-days":110,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Complexity"],"published-print":{"date-parts":[[2025,1]]},"abstract":"<jats:p>\n                    In deep learning, the choice of activation function plays a vital role in enhancing model performance. We propose AHerfReLU, a novel activation function that combines the rectified linear unit (ReLU) function with the error function (erf), complemented by a regularization term 1\/(1 +\n                    <jats:italic>x<\/jats:italic>\n                    <jats:sup>2<\/jats:sup>\n                    ), ensuring smooth gradients even for negative inputs. The function is zero centered, bounded below, and nonmonotonic, offering significant advantages over traditional activation functions like ReLU. We compare AHerfReLU with 10 adaptive activation functions and state\u2010of\u2010the\u2010art activation functions, including ReLU, Swish, and Mish. Experimental results show that replacing ReLU with AHerfReLU leads to 3.18% improvement in Top\u20101 accuracy on the LeNet network for the CIFAR100 dataset, 0.63% improvement on CIFAR10%, and 1.3% improvement in mean average precision (mAP) on the SSD300 model in the Pascal VOC dataset. Our results demonstrate that AHerfReLU enhances model performance, offering improved accuracy, loss reduction, and convergence stability. The function outperforms existing activation functions, providing a promising alternative for deep learning tasks.\n                  <\/jats:p>","DOI":"10.1155\/cplx\/8233876","type":"journal-article","created":{"date-parts":[[2025,4,22]],"date-time":"2025-04-22T02:04:15Z","timestamp":1745287455000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":8,"title":["AHerfReLU: A Novel Adaptive Activation Function Enhancing Deep Neural Network Performance"],"prefix":"10.1155","volume":"2025","author":[{"given":"Abaid","family":"Ullah","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-8553-2481","authenticated-orcid":false,"given":"Muhammad","family":"Imran","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-7805-5938","authenticated-orcid":false,"given":"Muhammad Abdul","family":"Basit","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Madeeha","family":"Tahir","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-8634-3440","authenticated-orcid":false,"given":"Jihad","family":"Younis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2025,4,21]]},"reference":[{"key":"e_1_2_10_1_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10044-024-01277-w"},{"key":"e_1_2_10_2_2","unstructured":"NwankpaC. IjomahW. GachaganA. andMarshallS. Activation Functions: Comparison of Trends in Practice and Research for Deep Learning 2018."},{"key":"e_1_2_10_3_2","unstructured":"NairV.andHintonG. E. Rectified Linear Units Improve Restricted Boltzmann Machines Proceedings of the 27th International Conference on Machine Learning August 2010 807\u2013814."},{"key":"e_1_2_10_4_2","article-title":"Rectifier Nonlinearities Improve Neural Network Acoustic Models","volume":"30","author":"Maas A. L.","year":"2013","journal-title":"Processing ICML"},{"key":"e_1_2_10_5_2","doi-asserted-by":"crossref","unstructured":"HeK. ZhangX. RenS. andSunJ. Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification Proceedings of the IEEE International Conference on Computer Vision April 2015 1026\u20131034.","DOI":"10.1109\/ICCV.2015.123"},{"key":"e_1_2_10_6_2","unstructured":"ClevertD. A. Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) 2015."},{"key":"e_1_2_10_7_2","unstructured":"CarlileB. DelamarterG. KinneyP. MartiA. andWhitneyB. Improving Deep Learning by Inverse Square Root Linear Units (ISRLUs) 2017."},{"key":"e_1_2_10_8_2","doi-asserted-by":"crossref","unstructured":"QiuS. XuX. andCaiB. FReLU: Flexible Rectified Linear Units for Improving Convolutional Neural Networks 2018 24th International Conference on Pattern Recognition (ICPR) June 2018 1223\u20131228 https:\/\/doi.org\/10.1109\/icpr.2018.8546022 2-s2.0-85059764645.","DOI":"10.1109\/ICPR.2018.8546022"},{"key":"e_1_2_10_9_2","volume-title":"Empirical Evaluation of Rectified Activations in Convolutional Network","author":"Xu B.","year":"2015"},{"key":"e_1_2_10_10_2","volume-title":"Searching for Activation Functions","author":"Ramachandran P.","year":"2017"},{"key":"e_1_2_10_11_2","doi-asserted-by":"publisher","DOI":"10.1109\/access.2021.3085855"},{"key":"e_1_2_10_12_2","doi-asserted-by":"crossref","unstructured":"SivriT. T. AkmanN. P. andBerkolA. Multiclass Classification Using Arctangent Activation Function and Its Variations 2022 14th International Conference on Electronics Computers and Artificial Intelligence (ECAI) March 2022 1\u20136 https:\/\/doi.org\/10.1109\/ecai54874.2022.9847486.","DOI":"10.1109\/ECAI54874.2022.9847486"},{"key":"e_1_2_10_13_2","doi-asserted-by":"crossref","unstructured":"LauM. M.andLimK. H. Review of Adaptive Activation Function in Deep Neural Network 2018 IEEE-EMBS Conference on Biomedical Engineering and Sciences (IECBES) September 2018 686\u2013690.","DOI":"10.1109\/IECBES.2018.8626714"},{"key":"e_1_2_10_14_2","doi-asserted-by":"crossref","unstructured":"KamalovF. NazirA. SafaralievM. CherukuriA. K. andZgheibR. Comparative Analysis of Activation Functions in Neural Networks 2021 28th IEEE International Conference on Electronics Circuits and Systems (ICECS) October 2021 1\u20136 https:\/\/doi.org\/10.1109\/icecs53924.2021.9665646.","DOI":"10.1109\/ICECS53924.2021.9665646"},{"key":"e_1_2_10_15_2","doi-asserted-by":"crossref","unstructured":"TrottierL. GiguereP. andChaib-draaB. Parametric Exponential Linear Unit for Deep Convolutional Neural Networks Processing 16th IEEE International Conference Machine Learning Application (ICMLA) December 2017 207\u2013214 https:\/\/doi.org\/10.1109\/icmla.2017.00038 2-s2.0-85048497850.","DOI":"10.1109\/ICMLA.2017.00038"},{"key":"e_1_2_10_16_2","doi-asserted-by":"crossref","unstructured":"ElfwingS. UchibeE. andDoyaK. Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning 2017.","DOI":"10.1016\/j.neunet.2017.12.012"},{"key":"e_1_2_10_17_2","volume-title":"Gaussian Error Linear Units (Gelus)","author":"Hendrycks D.","year":"2016"},{"key":"e_1_2_10_18_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-981-16-2248-9_34"},{"key":"e_1_2_10_19_2","doi-asserted-by":"crossref","unstructured":"RoyS. K. MannaS. DubeyS. R. andChaudhuriB. B. LiSHT: Non-parametric Linearly Scaled Hyperbolic Tangent Activation Function for Neural Networks International Conference on Computer Vision and Image Processing November 2022 462\u2013476 https:\/\/doi.org\/10.1007\/978-3-031-31407-0_35.","DOI":"10.1007\/978-3-031-31407-0_35"},{"key":"e_1_2_10_20_2","article-title":"Mish: A Self-Regularized Non-Monotonic Activation Function","author":"Misra D.","year":"2019","journal-title":"arXiv preprint arXiv:1908.08681"},{"key":"e_1_2_10_21_2","doi-asserted-by":"crossref","unstructured":"GodfreyL. B. An Evaluation of Parametric Activation Functions for Deep Learning 2019 IEEE International Conference on Systems Man and Cybernetics (SMC) October 2019 3006\u20133011.","DOI":"10.1109\/SMC.2019.8913972"},{"key":"e_1_2_10_22_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2022.01.001"},{"key":"e_1_2_10_23_2","doi-asserted-by":"publisher","DOI":"10.1109\/access.2021.3105355"},{"key":"e_1_2_10_24_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2022.117181"},{"key":"e_1_2_10_25_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2020.114534"},{"key":"e_1_2_10_26_2","doi-asserted-by":"crossref","unstructured":"NagS. BhattacharyyaM. MukherjeeA. andKunduR. SERF: Towards Better Training of Deep Neural Networks Using Log-Softplus Error Activation Function Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision April 2023 5324\u20135333.","DOI":"10.1109\/WACV56688.2023.00529"},{"key":"e_1_2_10_27_2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i6.20557"},{"key":"e_1_2_10_28_2","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1989.1.4.541"},{"key":"e_1_2_10_29_2","unstructured":"MolinaA. SchramowskiP. andKerstingK. Pad\\\u2032e activation units: End-to-end learning of flexible activation functions in deep networks 2019."},{"key":"e_1_2_10_30_2","unstructured":"KidgerP.andLyonsT. Universal Approximation with Deep Narrow Networks Conference on Learning Theory July 2020 PMLR 2306\u20132327."},{"key":"e_1_2_10_31_2","unstructured":"YannL. CorinnaC. andBurgesC. J. MNIST Handwritten Digit Database 2010."},{"key":"e_1_2_10_32_2","unstructured":"XiaoH. RasulK. andVollgrafR. Fashion-Mnist: A Novel Image Dataset for Benchmarking Machine Learning Algorithms 2017."},{"key":"e_1_2_10_33_2","article-title":"Reading Digits in Natural Images with Unsupervised Feature Learning","volume":"2011","author":"Netzer Y.","year":"2011","journal-title":"NIPS Workshop on Deep Learning and Unsupervised Feature Learning"},{"key":"e_1_2_10_34_2","unstructured":"KrizhevskyA.andHintonG. Learning Multiple Layers of Features from Tiny Images 2009."},{"key":"e_1_2_10_35_2","article-title":"Tiny Imagenet Visual Recognition Challenge","volume":"7","author":"Le Y.","year":"2015","journal-title":"CS 231N"},{"key":"e_1_2_10_36_2","unstructured":"IoffeS. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift 2015."},{"key":"e_1_2_10_37_2","first-page":"1929","article-title":"Dropout: A Simple Way to Prevent Neural Networks from Overfitting","volume":"15","author":"Srivastava N.","year":"2014","journal-title":"Journal of Machine Learning Research"},{"key":"e_1_2_10_38_2","unstructured":"SimonyanK. Very Deep Convolutional Networks for Large-Scale Image Recognition 2014."},{"key":"e_1_2_10_39_2","doi-asserted-by":"crossref","unstructured":"HeK. ZhangX. RenS. andSunJ. Deep Residual Learning for Image Recognition Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition August 2016 770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"key":"e_1_2_10_40_2","doi-asserted-by":"publisher","DOI":"10.5244\/C.30.87"},{"key":"e_1_2_10_41_2","doi-asserted-by":"publisher","DOI":"10.1109\/5.726791"},{"key":"e_1_2_10_42_2","doi-asserted-by":"crossref","unstructured":"HuangG. LiuZ. Van Der MaatenL. andWeinbergerK. Q. Densely Connected Convolutional Networks Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition April 2017 4700\u20134708.","DOI":"10.1109\/CVPR.2017.243"},{"key":"e_1_2_10_43_2","doi-asserted-by":"publisher","DOI":"10.1214\/aoms\/1177729586"},{"key":"e_1_2_10_44_2","doi-asserted-by":"crossref","unstructured":"CordtsM. OmranM. RamosS. RehfeldT. EnzweilerM. andBenensonR. The Cityscapes Dataset for Semantic Urban Scene Understanding Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition May 2016 3213\u20133223 https:\/\/doi.org\/10.1109\/cvpr.2016.350 2-s2.0-84986255616.","DOI":"10.1109\/CVPR.2016.350"},{"key":"e_1_2_10_45_2","doi-asserted-by":"crossref","unstructured":"RonnebergerO. FischerP. andBroxT. U-Net: Convolutional Networks for Biomedical Image Segmentation Medical Image Computing and Computer-Assisted Intervention\u2013MICCAI 2015: 18th International Conference October 2015.","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_2_10_46_2","article-title":"Adam: A Method for Stochastic Optimization","author":"Kingma D. P.","year":"2014","journal-title":"arXiv preprint arXiv:1412.6980"},{"key":"e_1_2_10_47_2","unstructured":"GlorotX.andBengioY. Understanding the Difficulty of Training Deep Feedforward Neural Networks Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics March 2010 JMLR Workshop and Conference Proceedings 249\u2013256."},{"key":"e_1_2_10_48_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-009-0275-4"},{"key":"e_1_2_10_49_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46448-0_2"},{"key":"e_1_2_10_50_2","article-title":"Attention Is All You Need","author":"Vaswani A.","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_10_51_2","doi-asserted-by":"publisher","DOI":"10.3390\/electronics11040540"}],"container-title":["Complexity"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1155\/cplx\/8233876","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1155\/cplx\/8233876","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1155\/cplx\/8233876","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,9]],"date-time":"2026-03-09T03:49:12Z","timestamp":1773028152000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1155\/cplx\/8233876"}},"subtitle":[],"editor":[{"given":"Jesus Manuel","family":"Munoz-Pacheco","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2025,1]]},"references-count":51,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2025,1]]}},"alternative-id":["10.1155\/cplx\/8233876"],"URL":"https:\/\/doi.org\/10.1155\/cplx\/8233876","archive":["Portico"],"relation":{},"ISSN":["1076-2787","1099-0526"],"issn-type":[{"value":"1076-2787","type":"print"},{"value":"1099-0526","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1]]},"assertion":[{"value":"2024-09-09","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-03-20","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-04-21","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"8233876"}}