Generating Sequences With Recurrent Neural Networks
This work addresses the problem of sequence generation for applications like text and handwriting synthesis, offering a method to produce complex, realistic outputs.
The paper tackled generating sequences with long-range structure using Long Short-term Memory recurrent neural networks, demonstrating the approach for text and online handwriting, and extended it to synthesize realistic cursive handwriting conditioned on text.
This paper shows how Long Short-term Memory recurrent neural networks can be used to generate complex sequences with long-range structure, simply by predicting one data point at a time. The approach is demonstrated for text (where the data are discrete) and online handwriting (where the data are real-valued). It is then extended to handwriting synthesis by allowing the network to condition its predictions on a text sequence. The resulting system is able to generate highly realistic cursive handwriting in a wide variety of styles.