Gated Word-Character Recurrent Language Model

Abstract

We introduce a recurrent neural network language model (RNN-LM) with long short-term memory (LSTM) units that utilizes both character-level and word-level inputs. Our model has a gate that adaptively finds the optimal mixture of the character-level and word-level inputs. The gate creates the final vector representation of a word by combining two distinct representations of the word. The character-level inputs are converted into vector representations of words using a bidirectional LSTM. The word-level inputs are projected into another high-dimensional space by a word lookup table. The final vector representations of words are used in the LSTM language model which predicts the next word given all the preceding words. Our model with the gating mechanism effectively utilizes the character-level inputs for rare and out-of-vocabulary words and outperforms word-level language models on several English corpora.

References in corpus (3)

• Recurrent Neural Network Regularization
• Exploring the Limits of Language Modeling
• Adaptive Joint Learning of Compositional and Non-Compositional Phrase Embeddings

Cited by in corpus (18)

• An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
• Trellis Networks for Sequence Modeling
• Do Convolutional Networks need to be Deep for Text Classification ?
• Attending to Characters in Neural Sequence Labeling Models
• A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Tasks
• A Survey on Neural Network Language Models
• Words or Characters? Fine-grained Gating for Reading Comprehension
• Subword-augmented Embedding for Cloze Reading Comprehension
• Character-Word LSTM Language Models
• Restoring ancient text using deep learning: a case study on Greek epigraphy
• Nonsymbolic Text Representation
• Recurrent Neural Network from Adder's Perspective: Carry-lookahead RNN
• Strawman: an Ensemble of Deep Bag-of-Ngrams for Sentiment Analysis
• Character-based Neural Networks for Sentence Pair Modeling
• Syllable-level Neural Language Model for Agglutinative Language
• Question-Aware Sentence Gating Networks for Question and Answering
• Subword Language Model for Query Auto-Completion
• An Embedded Deep Learning based Word Prediction