Why do we need recurrent neural network? How does back propagation learn to minimize the error? Explain

Standard feedforward neural networks (FNN) have a fundamental limitation: they assume all inputs are independent of each other. Each input is processed in isolation, and the network has no memory of previous inputs. This creates problems for sequential data, where the current output depends on pr...