Abstract:This paper examines system combination issue for Syllable-confusion-network (SCN) -based Chinese Spoken term detection (STD). System combination for STD usually leads to improved accuracy but suffers from increased index size or complicated index structure. But in the scenarios where the index size and search speed are critical, a single compact index is highly desirable. Therefore we explore methods for efficient combination of a word-based system and a syllable-based system while keeping the compactness of the indices. First, a composite SCN is generated using two approaches: lattice combination and confusion network combination. Then a simple compact index is constructed from this composite SCN by merging cross-system redundant information. The experimental result on a 60-hour corpus shows that a relative accuracy improvement of 16.20% is achieved over the baseline syllable-based system. Meanwhile, it reduces the index size by 22.3% compared to the commonly adopted score combination method under comparable accuracy.

Syllable Based Audio Search Using Confusion Network Arc As Indexing Unit

Fast Fuzzy Keyword Spotting Using Syllable Confusion Network Indexing

A Fast Fuzzy Keyword Spotting Algorithm Based on Syllable Confusion Network.

Keyword Spotting Based on Syllable Confusion Network.

A Novel Chinese Mandarin Speech Indexing Method Based on Confusion Network Using Tone Information

Efficient System Combination for Chinese Spoken Term Detection

Efficient System Combination for Syllable-Confusion-Network-Based Chinese Spoken Term Detection

Study on Chinese speech retrieval based on posterior probability

Improving the Syllable-Synchronous Network Search Algorithm for Word Decoding in Continuous Chinese Speech Recognition

Spoken Document Retrieval Based On Confusion Network With Syllable Fragments

Keyword Spotting Based on Phoneme Confusion Matrix

End-to-end keywords spotting based on connectionist temporal classification for Mandarin

Syllable Lattice Based Chinese Speech Retrieval Techniques and Removing Redundancy Method from Indices

Lattice-based continuous Chinese speech indexing based on mutual information confidence measure

Chinese Spoken Document Retrieval Based On Syllable Neighbor Posterior Probability Matrix

A Two-Step Keyword Spotting Method Using Fuzzy Search Algorithm

Lattice-based Indexing for Spontaneous Mandarin Speech

Deep neural networks for syllable based acoustic modeling in Chinese speech recognition.

Fusing Multiple Systems into a Compact Lattice Index for Chinese Spoken Term Detection.

EffectiveASR: A Single-Step Non-Autoregressive Mandarin Speech Recognition Architecture with High Accuracy and Inference Speed

Robust lyric search based on weighted syllable confusion matrix