SEARCH RESULT

This article explores the vital role that activation functions (AFs) play in deep learning and neural networks. AFs are essential elements that help hidden layers and the output layer communicate with one another. They are also crucial for controlling calculations and computations in these designs. One unique aspect of this survey is its thorough cataloguing of most AFs used in deep learning applications, along with an explanation of current trends in their real-world use. By methodically presenting the dynamic interaction between AF applications in real-world contexts and the most recent results from the deep learning literature, this endeavour stands out as a ground-breaking addition, setting it apart from traditional AF-centric research. This work is notable for its timing and for providing a novel investigation that surpasses previous AFfocused research. This study is a priceless tool for practitioners and academics, helping them make well-informed decisions by offering an unmatched synthesis of AF trends in real-world applications coupled with research findings. Beyond improving our knowledge of whether AI is appropriate for different applications, this paper creates, for the first time in the broad field of deep learning, a comprehensive compilation that clarifies the intimate relationship between AI applications and the state-of-the-art research in the field.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Jay Mehta Srushti.v.Vaidya