Author Identification using Multi-headed Recurrent Neural Networks
This addresses the problem of author identification for forensic or literary analysis, but it is incremental as it builds on existing RNN methods with a novel architectural tweak.
The paper tackled author identification with limited training data by splitting the output layer of a character-level RNN into author-specific sub-models while sharing the recurrent layer, achieving first place in two out of four languages in the PAN 2015 task.
Recurrent neural networks (RNNs) are very good at modelling the flow of text, but typically need to be trained on a far larger corpus than is available for the PAN 2015 Author Identification task. This paper describes a novel approach where the output layer of a character-level RNN language model is split into several independent predictive sub-models, each representing an author, while the recurrent layer is shared by all. This allows the recurrent layer to model the language as a whole without over-fitting, while the outputs select aspects of the underlying model that reflect their author's style. The method proves competitive, ranking first in two of the four languages.