Many real systems rely on sequential data, yet it is not always clear which deep learning model to use for forecasting or pattern prediction.
What you'll learn
Many real systems rely on sequential data, yet it is not always clear which deep learning model to use for forecasting or pattern prediction. In this course, Recurrent Neural Networks (RNNs), you'll gain the ability to model sequences with modern recurrent architectures. First, you'll explore how simple RNNs represent temporal information, and apply them to a real-time series forecasting problem. Next, you'll discover how LSTMs and GRUs improve training stability and accuracy through gating mechanisms. Finally, you'll learn how attention-based models extend these ideas and handle longer-range dependencies. When you're finished with this course, you'll have the skills and knowledge of recurrent sequence modeling needed to select, train, and evaluate models for your own sequential data tasks.