What is the exploding gradient problem in deep learning and why does it occur during training? How does it affect the performance and stability of neural networks? What are the common causes of exploding gradients in deep neural networks? What techniques can be used to prevent or control exploding gradients? How does it compare with the vanishing gradient problem in deep learning models?