{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/adaptive-hybrid-activation-function-for-deep","title":"Adaptive hybrid activation function for deep neural networks","arxiv_id":null,"date":"2022-04-25","proceeding":"System research and information technologies 2022 4","authors":["Yevgeniy Bodyanskiy","Serhii Kostiuk"],"abstract":"The adaptive hybrid activation function (AHAF) is proposed that combines the properties of the rectifier units and the squashing functions. The proposed function can be used as a drop-in replacement for ReLU, SiL and Swish activations for deep neural networks and can evolve to one of such functions during the training. The effectiveness of the function was evaluated on the image classification task using the Fashion-MNIST and CIFAR-10 datasets. The evaluation shows that the neural networks with AHAF activations achieve better classification accuracy comparing to their base implementations that use ReLU and SiL. A double-stage parameter tuning process for training the neural networks with AHAF is proposed. The proposed approach is sufficiently simple from the implementation standpoint and provides high performance for the neural network training process.","url_abs":"https://doi.org/10.20535/SRIT.2308-8893.2022.1.07","url_pdf":"http://journal.iasa.kpi.ua/article/download/259203/255848/596453","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"adaptive-hybrid-activation-function-for-deep","repo_url":"https://github.com/s-kostyuk/ahaf_activation_pytorch","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"activation-function-synthesis","task_name":"Activation Function Synthesis"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"learning-theory","task_name":"Learning Theory"},{"task_slug":"architecture-search","task_name":"Neural Architecture Search"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[{"method_slug":"ahaf","method_name":"AHAF"},{"method_slug":"base","method_name":"BASE"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dspt","method_name":"DSPT"},{"method_slug":"silu","method_name":"SiLU"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"}],"datasets_introduced":[],"methods_introduced":[{"slug":"ahaf","name":"AHAF","full_name":"Adaptive Hybrid Activation Function"},{"slug":"dspt","name":"DSPT","full_name":"double-stage parameter tuning"}],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}