Recurrent Neural Networks

Thirteen chapters summarize the design and application of recurrent neural networks (RNN), and exemplify current research ideas and challenges in this subfield of artificial neural network research and development. The first section concentrates on ideas for alternate designs and advances in theoretical aspects of RNNs. Some authors discuss aspects of improving RNN performance and connections with Bayesian analysis and knowledge representation. The second section looks at recent applications of RNNs, such as trajectories, control systems, robotics, and language learning. Architectures and learning techniques are addressed in every chapter.