The hyperbolic tangent (tanh) function is a non-linear activation function used in neural networks that maps real-valued inputs to the range (–1 to 1).
Uses in AI:
Commonly used in older recurrent neural networks (RNNs)
Helps center activations around zero, improving convergence during training
Sometimes used in hidden layers when symmetry around zero is desired
The tanh function is a smooth, symmetric activation function that outputs values between –1 and 1, useful for zero-centered data, but less common in modern deep networks due to gradient issues.
