# Investigation on N-gram Approximated RNNLMs for Recognition of Morphologically Rich Speech

## TL;DR

This paper explores methods to reduce the complexity of RNN language models for recognizing morphologically rich Hungarian speech, achieving near-RNN performance with lower delay and improved accuracy.

## Contribution

It introduces RNN-BNLM approximation and morph-based modeling to reduce RNNLM complexity while maintaining accuracy in speech recognition.

## Key findings

- RNN-BNLM recovers 40% of RNN perplexity reduction
- Combining morph-based models with RNN approximation yields 8% WER reduction
- Real-time speech recognition is preserved with proposed methods

## Abstract

Recognition of Hungarian conversational telephone speech is challenging due to the informal style and morphological richness of the language. Recurrent Neural Network Language Model (RNNLM) can provide remedy for the high perplexity of the task; however, two-pass decoding introduces a considerable processing delay. In order to eliminate this delay we investigate approaches aiming at the complexity reduction of RNNLM, while preserving its accuracy. We compare the performance of conventional back-off n-gram language models (BNLM), BNLM approximation of RNNLMs (RNN-BNLM) and RNN n-grams in terms of perplexity and word error rate (WER). Morphological richness is often addressed by using statistically derived subwords - morphs - in the language models, hence our investigations are extended to morph-based models, as well. We found that using RNN-BNLMs 40% of the RNNLM perplexity reduction can be recovered, which is roughly equal to the performance of a RNN 4-gram model. Combining morph-based modeling and approximation of RNNLM, we were able to achieve 8% relative WER reduction and preserve real-time operation of our conversational telephone speech recognition system.