MEGA Hub
New

MERaLiON-GR: Speech Gender Recognition Model for English and SEA Languages

Published on Aug 5cs.CL

Authors

Do you know Qiongqiong Wang?You can claim authorship or link another user.Do you know Ai Ti Aw?You can claim authorship or link another user.Do you know Nancy F. Chen?You can claim authorship or link another user.Do you know Ying Lay Chiu?You can claim authorship or link another user.Do you know Yang Ding?You can claim authorship or link another user.Do you know Yingxu He?You can claim authorship or link another user.Do you know Ridong Jiang?You can claim authorship or link another user.Do you know Zhuohan Liu?You can claim authorship or link another user.Do you know Yanfeng Lu?You can claim authorship or link another user.Do you know Yi Ma?You can claim authorship or link another user.Do you know Muhammad Huzaifah?You can claim authorship or link another user.Do you know Nabilah Binte Md Johan?You can claim authorship or link another user.Do you know Nattadaporn Lertcheva?You can claim authorship or link another user.Do you know Pham Minh Duc?You can claim authorship or link another user.Do you know Sailor Hardik Bhupendra?You can claim authorship or link another user.Do you know Siti Umairah Binte Mohammad Salleh?You can claim authorship or link another user.Do you know Shuo Sun?You can claim authorship or link another user.Do you know Tarun Kumar Vangani?You can claim authorship or link another user.Do you know Jeremy H. M. Wong?You can claim authorship or link another user.Do you know Jinyang Wu?You can claim authorship or link another user.Do you know Longyin Zhang?You can claim authorship or link another user.

Abstract

We present MERaLiON-GR, a speech gender recognition system that performs binary classification (female / male) on English and Southeast Asian (SEA) languages. The model finetunes MERaLiON-SpeechEncoder-2, a large conformer based transformer pre-trained on a broad speech corpus, and applies parameter efficient fine-tuning via Low-Rank Adaptation (LoRA) to adapt the encoder to the gender recognition task, and appends a multi-scale ECAPA-TDNN down stream network with attention pooling and a lightweight linear classifier. Extensive evaluations across multilingual Singaporean and Southeast Asian languages (English, Chinese, Malay, Tamil, Thai, Vietnamese, Indonesian, and Khmer) show that MERaLiON-GR consistently surpasses the state-of-the-art gender recognition model Vox-Profile and a large Audio-LLM, in both full-utterance and segment level evaluation modes. The results underscore the value of dedicated speech models in achieving accurate paralinguistic understanding and strong cross-lingual generalization.

Community

00