Additive Phoneme-aware Margin Softmax Loss for Language Recognition

Abstract

This paper proposes an additive phoneme-aware margin softmax (APM-Softmax) loss to train the multi-task learning network with phonetic information for language recognition. In additive margin softmax (AM-Softmax) loss, the margin is set as a constant during the entire training for all training samples, and that is a suboptimal method since the recognition difficulty varies in training samples. In additive angular margin softmax (AAM-Softmax) loss, the additional angular margin is set as a costant as well. In this paper, we propose an APM-Softmax loss for language recognition with phoneitc multi-task learning, in which the additive phoneme-aware margin is automatically tuned for different training samples. More specifically, the margin of language recognition is adjusted according to the results of phoneme recognition. Experiments are reported on Oriental Language Recognition (OLR) datasets, and the proposed method improves AM-Softmax loss and AAM-Softmax loss in different language recognition testing conditions.

arXiv commentAccepted by Interspeech 2021

References in corpus (6)

• Deep Speaker: an End-to-End Neural Speaker Embedding System
• THCHS-30 : A Free Chinese Speech Corpus
• Convolutional Neural Networks and x-vector Embedding for DCASE2018 Acoustic Scene Classification Challenge
• Large Margin Softmax Loss for Speaker Verification
• Text adaptation for speaker verification with speaker-text factorized embeddings
• AP16-OL7: A Multilingual Database for Oriental Languages and A Language Recognition Baseline