Exploring the Limits of Language Modeling
This work addresses language understanding for NLP researchers by improving state-of-the-art performance on a benchmark, though it is incremental in nature.
The authors tackled large-scale language modeling by extending recurrent neural networks to handle large corpora and long-term language structure, achieving a perplexity reduction from 51.3 to 30.0 with fewer parameters and setting a new ensemble record of 23.7.
In this work we explore recent advances in Recurrent Neural Networks for large scale Language Modeling, a task central to language understanding. We extend current models to deal with two key challenges present in this task: corpora and vocabulary sizes, and complex, long term structure of language. We perform an exhaustive study on techniques such as character Convolutional Neural Networks or Long-Short Term Memory, on the One Billion Word Benchmark. Our best single model significantly improves state-of-the-art perplexity from 51.3 down to 30.0 (whilst reducing the number of parameters by a factor of 20), while an ensemble of models sets a new record by improving perplexity from 41.0 down to 23.7. We also release these models for the NLP and ML community to study and improve upon.