Bin Wang
 |
Scientist
|
News
2025.01 Our AudioBench work is officially accepted to NAACL 2025 Main Conference! Leaderboard and Code.
2024.12 We released MERaLiON models, the first audio-based large language model designed specifically for Singlish and other related tasks!
2024-2026 Our team is working on National LLM Project Press Coverage.
2024 AudioBench is released with leaderboard and evaluation toolkit - AudioLLM Evaluation.
2024 NAACL SeaEval for Multilingual Evaluation.
About Me
I am a scientist at Aural & Language Intelligence Department, I2R, A*STAR. Before joing that, I was a research fellow at National University of Singapore (NUS) working with Prof. Haizhou Li from 2021-2023. I received my Ph.D. degree from University of Southern California (USC) supervised by Prof. C.-C. Jay Kuo in 2021. My bachelor's degree is obtained from University of Electronic Science and Technology of China (UESTC) in 2017.
Some of the topics that I am currently researching include:
Make LLM can hear - AudioLLM - Audio-Based Large Language Models
What techniques can be used to effectively integrate audio processing capabilities into existing LLM architectures?
What is the most efficient approach for achieving seamless cross-modality integration?
What benchmarks can be designed to accurately evaluate the real-world performance of AudioLLMs?
Current Outcomes: MERaLiON-AduioLLM, AudioBench, Awesome-Audio-LLM, MoWE-Audio
Multilingal and Multicultual LLM
What unique properties should a multilingual LLM possess to cater to diverse languages effectively?
How can multilingual learning be made more efficient and effective, especially for low-resource languages?
What internal mechanisms can ensure robust multilingual knowledge alignment within the model?
Conversional AI
Representation Learning for Retrieval-Augmented Generation, Knowledge Graphs
What representation and coordination strategies can enhance multi-agent communication in shared environments?
What methods can enable conversational agents to effectively reason and plan based on learned or provided world models?
Current Outcomes: Representation Learning, Commonsense Knowledge Graph
Opportunities
We are actively looking for candidates working on Multimodal LLMs (text, audio, vision, etc.).
Research Interns (6 months or above preferred),
Ph.D. Students
Long-term Positions
Research Engineer / Scientist (Both engineering and research background are preferred.)
Some Publications
Bin Wang, Xunlong Zou, Geyu Lin, Shuo Sun, Zhuohan Liu, Wenyu Zhang, Zhengyuan Liu, AiTi Aw, Nancy F. Chen. “AudioBench: A Universal Benchmark for Audio Large Language Models.” NAACL, 2025. [paper], [code]
Bin Wang, Zhengyuan Liu, Xin Huang, Fangkai Jiao, Yang Ding, AiTi Aw, Nancy F. Chen. “SeaEval for Multilingual Foundation Models: From Cross-Lingual Alignment to Cultural Reasoning.” NAACL, 2024. [paper], [code]
Bin Wang, Chen Zhang, Yan Zhang, Yiming Chen and Haizhou Li. “Analyzing and Evaluating Faithfulness in Dialogue Summarization.” EMNLP, 2022. [paper], [code]
Bin Wang, C.-C. Jay Kuo, and Haizhou Li. “Just Rank: Rethinking Evaluation with Word and Sentence Similarities.” ACL, 2022. [paper], [code]
Bin Wang, Guangtao Wang, Jing Huang, Jiaxuan You, Jure Leskovec, and C.-C. Jay Kuo. “Inductive learning on commonsense knowledge graph completion.” IJCNN, 2021. [paper], [code]
Bin Wang, and C.-C. Jay Kuo. “SBERT-WK: A sentence embedding method by dissecting bert-based word models.” IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2020. [paper], [code]
Bin Wang*, Angela Wang*, Fenxiao Chen, Yuncheng Wang, and C.-C. Jay Kuo. “Evaluating word embedding models: methods and experimental results.” APSIPA transactions on signal and information processing, 2019. [paper], [code]