Implementing contextual biasing in GPU decoder for online ASR

TL;DR

This paper introduces a method for integrating contextual biasing into real-time GPU decoding for online ASR, enabling dynamic rescoring and improving prediction accuracy without lattice generation.

Contribution

It presents a novel approach to implement contextual biasing directly in GPU decoding for online ASR, including dynamic context switching and compatibility with Kaldi GPU decoder.

Findings

Effective biasing of partial ASR predictions on GPU

Supports dynamic context switching during decoding

Code is publicly available and tested on open datasets

Abstract

GPU decoding significantly accelerates the output of ASR predictions. While GPUs are already being used for online ASR decoding, post-processing and rescoring on GPUs have not been properly investigated yet. Rescoring with available contextual information can considerably improve ASR predictions. Previous studies have proven the viability of lattice rescoring in decoding and biasing language model (LM) weights in offline and online CPU scenarios. In real-time GPU decoding, partial recognition hypotheses are produced without lattice generation, which makes the implementation of biasing more complex. The paper proposes and describes an approach to integrate contextual biasing in real-time GPU decoding while exploiting the standard Kaldi GPU decoder. Besides the biasing of partial ASR predictions, our approach also permits dynamic context switching allowing a flexible rescoring per each speech segment directly on GPU. The code is publicly released and tested with open-sourced test sets.