System for fast lexical and phonetic spoken term detection in a Czech cultural heritage archive

Abstract

The main objective of the work presented in this paper was to develop a complete system that would accomplish the original visions of the MALACH project. Those goals were to employ automatic speech recognition and information retrieval techniques to provide improved access to the large video archive containing recorded testimonies of the Holocaust survivors. The system has been so far developed for the Czech part of the archive only. It takes advantage of the state-of-the-art speech recognition system tailored to the challenging properties of the recordings in the archive (elderly speakers, spontaneous speech and emotionally loaded content) and its close coupling with the actual search engine. The design of the algorithm adopting the spoken term detection approach is focused on the speed of the retrieval. The resulting system is able to search through the 1,000 h of video constituting the Czech portion of the archive and find query word occurrences in the matter of seconds. The phonetic search implemented alongside the search based on the lexicon words allows to find even the words outside the ASR system lexicon such as names, geographic locations or Jewish slang.

A Relevance Score Estimation for Spoken Term Detection Based on RNN-Generated Pronunciation Embeddings

“…First of all, we used the baseline empirical approach to phoneme-based search described in [10]. The first three rows of Tab.…”

“…In this paper, we use the index-based method for determining the occurrence candidates. The method is based on the empirical approach [10], but it is used only to determine the candidates and no relevance score estimation is performed. To index the archive we use the sub-word units, particularly the triplets of phonemes.…”

“…Our method is designed for the use in the STD task with large spoken archives in mind. It is a machine learning based extension of an empirical method described in [10]. The spoken archive is processed off-line, i.e.…”

Dynamic Threshold Selection Method for Multi-label Newspaper Topic Identification

“…Lemmatization has been shown to improve the results when dealing with sparse data in the area of information retrieval [15] and spoken term detection [16] in highly inflected languages, on that account the experiments on the effects of lemmatization in the field of topic identification was performed [17]. As a result of these experiments the automatic text lemmatization is also applied in our work.…”

Application of Lemmatization and Summarization Methods in Topic Identification Module for Large Scale Language Modeling Data Filtering

“…Lemmatization has been shown to improve the results when dealing with sparse data in the area of information retrieval [4] and spoken term detection [10] in highly inflected languages, therefore the effects of lemmatization on topic identification accuracy is studied in the paper. On the other hand, since the system is used for processing large amounts of data, a summarization method was implemented and the effect of using only the summary of an article on the topic identification accuracy is studied.…”