Year: 2025
Type: preprint
Abstract: We present MERaLiON-SER, a robust speech emotion recognition model designed for English and Southeast Asian languages. The model is trained using a hybrid objective combining weighted categorical cross-entropy and Concordance Correlation Coefficient (CCC) losses for joint discrete and dimensional emotion modelling. This dual approach enables the model to capture both the distinct categories of emotion (like happy or angry) and the fine-grained, such as arousal (intensity), valence (positivity/negativity), and dominance (sense of control), leading to a more comprehensive and robust representation of human affect. Extensive evaluations across multilingual Singaporean languages (English, Chinese, Malay, and Tamil ) and other public benchmarks show that MERaLiON-SER consistently surpasses both open-source speech encoders and large Audio-LLMs. These results underscore the importance of specialised speech-only models for accurate paralinguistic understanding and cross-lingual generalisation. Furthermore, the proposed framework provides a foundation for integrating emotion-aware perception into future agentic audio systems, enabling more empathetic and contextually adaptive multimodal reasoning.
Source: arXiv (Cornell University)
Authors: Sailor, Hardik B., Ti, Aw Ai, Nancy, Chen Fang Yih, Lay, Chiu Ying, Yang, Ding, Yingxu, He, Ridong, Jiang, Jingtao, Li, Jingyi, Liao, Zhuohan, Liu, Yanfeng, Lu, Yi, Ma, Gupta, Manas, Shahrin, Muhammad Huzaifah Bin Md, Johan, Nabilah Binte Md, Lertcheva, Nattadaporn, Chunlei, Pan, Duc, Pham Minh, Subaidi, Siti Maryam Binte Ahmad, Salleh, Siti Umairah Binte Mohammad, Shuo, Sun, Vangani, Tarun Kumar, Qiongqiong, Wang, Lewis, Won Cheng Yi, Jeremy, Wong Heng Meng, Jinyang, Wu, Huayun, Zhang, Longyin, Zhang, Xunlong, Zou
Language: English
Topic: Emotion and Mood Recognition
Subfield: Experimental and Cognitive Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W4416222584