Výsledky vyhledávání - "Styles, Suzy J."

Report

Investigating model performance in language identification: beyond simple error statistics

Autor: Styles, Suzy J., Chua, Victoria Y. H., Woon, Fei Ting, Liu, Hexin, Perera, Leibny Paola Garcia, Khudanpur, Sanjeev, Khong, Andy W. H., Dauwels, Justin

Language development experts need tools that can automatically identify languages from fluent, conversational speech, and provide reliable estimates of usage rates at the level of an individual recording. However, language identification systems are

Externí odkaz: http://arxiv.org/abs/2305.18925

Zobrazit plný text záznamu

Report

MERLIon CCS Challenge: A English-Mandarin code-switching child-directed speech corpus for language identification and diarization

Autor: Chua, Victoria Y. H., Liu, Hexin, Perera, Leibny Paola Garcia, Woon, Fei Ting, Wong, Jinyi, Zhang, Xiangyu, Khudanpur, Sanjeev, Khong, Andy W. H., Dauwels, Justin, Styles, Suzy J.

To enhance the reliability and robustness of language identification (LID) and language diarization (LD) systems for heterogeneous populations and scenarios, there is a need for speech processing models to be trained on datasets that feature diverse

Externí odkaz: http://arxiv.org/abs/2305.18881

Zobrazit plný text záznamu

Report

MERLIon CCS Challenge Evaluation Plan

Autor: Perera, Leibny Paola Garcia, Chua, Y. H. Victoria, Liu, Hexin, Woon, Fei Ting, Khong, Andy W. H., Dauwels, Justin, Khudanpur, Sanjeev, Styles, Suzy J.

This paper introduces the inaugural Multilingual Everyday Recordings- Language Identification on Code-Switched Child-Directed Speech (MERLIon CCS) Challenge, focused on developing robust language identification and language diarization systems that a

Externí odkaz: http://arxiv.org/abs/2305.19493

Zobrazit plný text záznamu

Report

PHO-LID: A Unified Model Incorporating Acoustic-Phonetic and Phonotactic Information for Language Identification

Autor: Liu, Hexin, Perera, Leibny Paola Garcia, Khong, Andy W. H., Styles, Suzy J., Khudanpur, Sanjeev

We propose a novel model to hierarchically incorporate phoneme and phonotactic information for language identification (LID) without requiring phoneme annotations for training. In this model, named PHO-LID, a self-supervised phoneme segmentation task

Externí odkaz: http://arxiv.org/abs/2203.12366

Zobrazit plný text záznamu

Report

Enhance Language Identification using Dual-mode Model with Knowledge Distillation

Autor: Liu, Hexin, Perera, Leibny Paola Garcia, Khong, Andy W. H., Dauwels, Justin, Styles, Suzy J., Khudanpur, Sanjeev

In this paper, we propose to employ a dual-mode framework on the x-vector self-attention (XSA-LID) model with knowledge distillation (KD) to enhance its language identification (LID) performance for both long and short utterances. The dual-mode XSA-L

Externí odkaz: http://arxiv.org/abs/2203.03218

Zobrazit plný text záznamu