Fast and Accurate OOV Decoder on High-Level Features

TL;DR

This paper introduces a fast, efficient OOV keyword search method using high-level phoneme posterior features from ASR systems, outperforming existing approaches in speed and accuracy on Georgian language data.

Contribution

The paper presents a novel, simple, and fast OOV decoder based on phoneme posterior features, significantly improving performance over state-of-the-art proxy-based methods.

Findings

Outperforms state-of-the-art in MTWV metric and speed

Demonstrates effectiveness on Georgian language data

Offers low memory usage and easy implementation

Abstract

This work proposes a novel approach to out-of-vocabulary (OOV) keyword search (KWS) task. The proposed approach is based on using high-level features from an automatic speech recognition (ASR) system, so called phoneme posterior based (PPB) features, for decoding. These features are obtained by calculating time-dependent phoneme posterior probabilities from word lattices, followed by their smoothing. For the PPB features we developed a special novel very fast, simple and efficient OOV decoder. Experimental results are presented on the Georgian language from the IARPA Babel Program, which was the test language in the OpenKWS 2016 evaluation campaign. The results show that in terms of maximum term weighted value (MTWV) metric and computational speed, for single ASR systems, the proposed approach significantly outperforms the state-of-the-art approach based on using in-vocabulary proxies for OOV keywords in the indexed database. The comparison of the two OOV KWS approaches on the fusion results of the nine different ASR systems demonstrates that the proposed OOV decoder outperforms the proxy-based approach in terms of MTWV metric given the comparable processing speed. Other important advantages of the OOV decoder include extremely low memory consumption and simplicity of its implementation and parameter optimization.