Minimising Biasing Word Errors for Contextual ASR with the Tree-Constrained Pointer Generator

TL;DR

This paper introduces TCPGen, a tree-constrained pointer generator for end-to-end ASR that effectively biases recognition towards long-tail words using external context, improving accuracy with minimal overhead.

Contribution

The paper presents a novel TCPGen component and associated training and testing methods that enhance biasing word recognition in ASR systems, especially for out-of-vocabulary words.

Findings

Significant WER reductions on biasing words, up to 40% relative improvement.

TCPGen enables zero-shot learning of unseen biasing words.

Enhanced biasing word recognition with minimal computational overhead.

Abstract

Contextual knowledge is essential for reducing speech recognition errors on high-valued long-tail words. This paper proposes a novel tree-constrained pointer generator (TCPGen) component that enables end-to-end ASR models to bias towards a list of long-tail words obtained using external contextual information. With only a small overhead in memory use and computation cost, TCPGen can structure thousands of biasing words efficiently into a symbolic prefix-tree and creates a neural shortcut between the tree and the final ASR output to facilitate the recognition of the biasing words. To enhance TCPGen, we further propose a novel minimum biasing word error (MBWE) loss that directly optimises biasing word errors during training, along with a biasing-word-driven language model discounting (BLMD) method during the test. All contextual ASR systems were evaluated on the public Librispeech audiobook corpus and the data from the dialogue state tracking challenges (DSTC) with the biasing lists extracted from the dialogue-system ontology. Consistent word error rate (WER) reductions were achieved with TCPGen, which were particularly significant on the biasing words with around 40\% relative reductions in the recognition error rates. MBWE and BLMD further improved the effectiveness of TCPGen and achieved more significant WER reductions on the biasing words. TCPGen also achieved zero-shot learning of words not in the audio training set with large WER reductions on the out-of-vocabulary words in the biasing list.