Attentive Temporal Pooling for Conformer-based Streaming Language Identification in Long-form Speech

Quan Wang, Yang Yu, Jason Pelecanos, Yiling Huang, Ignacio Lopez Moreno

Abstract

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.
BibTeX
@misc{wang2022attentive,
  author = {Quan Wang and Yang Yu and Jason Pelecanos and Yiling Huang and Ignacio Lopez Moreno},
  title = {{Attentive Temporal Pooling for Conformer-based Streaming Language Identification in Long-form Speech}},
  year = {2022},
  eprint = {2202.12163},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2202.12163},
}

Loading related content...

Figure