In this paper, we introduce a novel language identification system based on conformer layers. We propose an attentive temporal pooling mechanism to allow the model to carry information in long-form audio via a recurrent form, such that the inference can be performed in a streaming fashion. Additionally, we investigate two domain adaptation approaches to allow adapting an existing language identification model without retraining the model parameters for a new domain. We perform a comparative study of different model topologies under different constraints of model size, and find that conformer-based models significantly outperform LSTM and transformer based models. Our experiments also show that attentive temporal pooling and domain adaptation improve model accuracy.
@article{arxiv.2202.12163,
title = {Attentive Temporal Pooling for Conformer-based Streaming Language Identification in Long-form Speech},
author = {Quan Wang and Yang Yu and Jason Pelecanos and Yiling Huang and Ignacio Lopez Moreno},
journal= {arXiv preprint arXiv:2202.12163},
year = {2022}
}