# Spoken Term Detection and Relevance Score Estimation using Dot-Product of Pronunciation Embeddings

arXiv: https://arxiv.org/abs/2210.11895 · HTML: https://mlsanity.com/paper/2210.11895/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 24 references that are in the map.
- First posted on arXiv 21 Oct 2022.

## Builds on (most cited first, 5 of 5)

- [Deep convolutional acoustic word embeddings using word-pair side information](https://mlsanity.com/paper/1510.01032/) (2015)
- [Discriminative Acoustic Word Embeddings: Recurrent Neural Network-Based Approaches](https://mlsanity.com/paper/1611.02550/) (2016)
- [Multi-view Recurrent Neural Acoustic Word Embeddings](https://mlsanity.com/paper/1611.04496/) (2016)
- [Segmental Audio Word2Vec: Representing Utterances as Sequences of Vectors with Applications in Spoken Term Detection](https://mlsanity.com/paper/1808.02228/) (2018)
- [Wake Word Detection with Alignment-Free Lattice-Free MMI](https://mlsanity.com/paper/2005.08347/) (2020)

## Built on by (most cited first, 5 of 5)

- [Pretraining End-to-End Keyword Search with Automatically Discovered Acoustic Units](https://mlsanity.com/paper/2407.04652/) (2024)
- [Transformer-based encoder-encoder architecture for Spoken Term Detection](https://mlsanity.com/paper/2211.01089/) (2022)
- [End-to-End Open Vocabulary Keyword Search With Multilingual Neural Representations](https://mlsanity.com/paper/2308.08027/) (2023)
- [Written Term Detection Improves Spoken Term Detection](https://mlsanity.com/paper/2407.04601/) (2024)
- [PatchDSU: Uncertainty Modeling for Out of Distribution Generalization in Keyword Spotting](https://mlsanity.com/paper/2508.03190/) (2025)

## Abstract (arXiv)

The paper describes a novel approach to Spoken Term Detection (STD) in large spoken archives using deep LSTM networks. The work is based on the previous approach of using Siamese neural networks for STD and naturally extends it to directly localize a spoken term and estimate its relevance score. The phoneme confusion network generated by a phoneme recognizer is processed by the deep LSTM network which projects each segment of the confusion network into an embedding space. The searched term is projected into the same embedding space using another deep LSTM network. The relevance score is then computed using a simple dot-product in the embedding space and calibrated using a sigmoid function to predict the probability of occurrence. The location of the searched term is then estimated from the sequence of output probabilities. The deep LSTM networks are trained in a self-supervised manner from paired recognition hypotheses on word and phoneme levels. The method is experimentally evaluated on MALACH data in English and Czech languages.