Natural Language Statistical Features of LSTM-generated Texts
This work provides quantitative evaluation of LSTM text generation for researchers in NLP, but is incremental as it builds on existing analysis methods.
The study compared statistical features of LSTM-generated texts to human-written language and Markov models, finding that LSTMs reproduce long-range correlations similar to natural language and have an optimal temperature parameter for closeness to real texts.
Long Short-Term Memory (LSTM) networks have recently shown remarkable performance in several tasks dealing with natural language generation, such as image captioning or poetry composition. Yet, only few works have analyzed text generated by LSTMs in order to quantitatively evaluate to which extent such artificial texts resemble those generated by humans. We compared the statistical structure of LSTM-generated language to that of written natural language, and to those produced by Markov models of various orders. In particular, we characterized the statistical structure of language by assessing word-frequency statistics, long-range correlations, and entropy measures. Our main finding is that while both LSTM and Markov-generated texts can exhibit features similar to real ones in their word-frequency statistics and entropy measures, LSTM-texts are shown to reproduce long-range correlations at scales comparable to those found in natural language. Moreover, for LSTM networks a temperature-like parameter controlling the generation process shows an optimal value---for which the produced texts are closest to real language---consistent across all the different statistical features investigated.