CTC-aligned Audio-Text Embedding for Streaming Open-vocabulary Keyword Spotting

TL;DR

This paper presents a streaming open-vocabulary keyword spotting method that dynamically aligns audio and text embeddings using CTC, achieving competitive accuracy with low latency and model complexity.

Contribution

It introduces the first dynamic CTC-based alignment approach for joint audio-text embedding in streaming keyword spotting, enabling real-time open-vocabulary detection.

Findings

Achieves competitive performance on LibriPhrase dataset.

Uses only 155K model parameters.

Operates with O(U) decoding complexity.

Abstract

This paper introduces a novel approach for streaming openvocabulary keyword spotting (KWS) with text-based keyword enrollment. For every input frame, the proposed method finds the optimal alignment ending at the frame using connectionist temporal classification (CTC) and aggregates the frame-level acoustic embedding (AE) to obtain higher-level (i.e., character, word, or phrase) AE that aligns with the text embedding (TE) of the target keyword text. After that, we calculate the similarity of the aggregated AE and the TE. To the best of our knowledge, this is the first attempt to dynamically align the audio and the keyword text on-the-fly to attain the joint audio-text embedding for KWS. Despite operating in a streaming fashion, our approach achieves competitive performance on the LibriPhrase dataset compared to the non-streaming methods with a mere 155K model parameters and a decoding algorithm with time complexity O(U), where U is the length of the target keyword at inference time.