{"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/the-quest-for-the-golden-activation-function","title":"The Quest for the Golden Activation Function","arxiv_id":"1808.00783","date":"2018-08-02","proceeding":null,"authors":["Mina Basirat","Peter M. Roth"],"abstract":"Deep Neural Networks have been shown to be beneficial for a variety of tasks,\nin particular allowing for end-to-end learning and reducing the requirement for\nmanual design decisions. However, still many parameters have to be chosen in\nadvance, also raising the need to optimize them. One important, but often\nignored system parameter is the selection of a proper activation function.\nThus, in this paper we target to demonstrate the importance of activation\nfunctions in general and show that for different tasks different activation\nfunctions might be meaningful. To avoid the manual design or selection of\nactivation functions, we build on the idea of genetic algorithms to learn the\nbest activation function for a given task. In addition, we introduce two new\nactivation functions, ELiSH and HardELiSH, which can easily be incorporated in\nour framework. In this way, we demonstrate for three different image\nclassification benchmarks that different activation functions are learned, also\nshowing improved results compared to typically used baselines.","url_abs":"http://arxiv.org/abs/1808.00783v1","url_pdf":"http://arxiv.org/pdf/1808.00783v1.pdf","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":[],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[{"method_slug":"elish","method_name":"ELiSH"},{"method_slug":"hardelish","method_name":"HardELiSH"}],"datasets_introduced":[],"methods_introduced":[{"slug":"elish","name":"ELiSH","full_name":"Exponential Linear Squashing Activation"},{"slug":"hardelish","name":"HardELiSH","full_name":"HardELiSH"}],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1808.00783","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}