Activations Through Extensions: A Framework To Boost Performance Of Neural Networks
This work addresses the challenge of enhancing neural network efficiency and effectiveness for practitioners, though it appears incremental as it builds upon and unifies existing activation function research.
The authors tackled the problem of improving neural network performance by proposing a framework that unifies and explains existing activation function works, and introduced novel techniques to create 'extensions' of neural networks through activation function operations. They theoretically and empirically demonstrated that these extensions yield performance benefits with minimal space and time complexity costs on standard test functions and real-world time-series datasets.
Activation functions are non-linearities in neural networks that allow them to learn complex mapping between inputs and outputs. Typical choices for activation functions are ReLU, Tanh, Sigmoid etc., where the choice generally depends on the application domain. In this work, we propose a framework/strategy that unifies several works on activation functions and theoretically explains the performance benefits of these works. We also propose novel techniques that originate from the framework and allow us to obtain ``extensions'' (i.e. special generalizations of a given neural network) of neural networks through operations on activation functions. We theoretically and empirically show that ``extensions'' of neural networks have performance benefits compared to vanilla neural networks with insignificant space and time complexity costs on standard test functions. We also show the benefits of neural network ``extensions'' in the time-series domain on real-world datasets.