# Knowledge Distillation For Recurrent Neural Network Language Modeling With Trust Regularization

arXiv: https://arxiv.org/abs/1904.04163 · HTML: https://mlsanity.com/paper/1904.04163/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Apr 2019.

## Built on by (most cited first, 5 of 5)

- [Preventing Catastrophic Forgetting and Distribution Mismatch in Knowledge Distillation via Synthetic Data](https://mlsanity.com/paper/2108.05698/) (2021)
- [Continual Learning for Fake Audio Detection](https://mlsanity.com/paper/2104.07286/) (2021)
- [Unsupervised Domain Expansion for Visual Categorization](https://mlsanity.com/paper/2104.00233/) (2021)
- [Spectral Pruning for Recurrent Neural Networks](https://mlsanity.com/paper/2105.10832/) (2021)
- [Knowledge Distillation Decision Tree for Unravelling Black-box Machine Learning Models](https://mlsanity.com/paper/2206.04661/) (2022)

## Abstract (arXiv)

Recurrent Neural Networks (RNNs) have dominated language modeling because of their superior performance over traditional N-gram based models. In many applications, a large Recurrent Neural Network language model (RNNLM) or an ensemble of several RNNLMs is used. These models have large memory footprints and require heavy computation. In this paper, we examine the effect of applying knowledge distillation in reducing the model size for RNNLMs. In addition, we propose a trust regularization method to improve the knowledge distillation training for RNNLMs. Using knowledge distillation with trust regularization, we reduce the parameter size to a third of that of the previously published best model while maintaining the state-of-the-art perplexity result on Penn Treebank data. In a speech recognition N-bestrescoring task, we reduce the RNNLM model size to 18.5% of the baseline system, with no degradation in word error rate(WER) performance on Wall Street Journal data set.