const research = [

Speech Processing and Machine Learning Laboratory (SPMLLab), NTU

type: Undergraduate Researcher period: Aug 2023 - Present Taipei, Taiwan Advisor: Prof. Hung-yi Lee

Introduced BILLY, merging persona vectors of LLMs to enhance the creativity and replace discussion of multi-LLM systems.
Boosting Originality by 15% and reducing the cost by 95% than multi-LLM baseline (Co-advised by Prof. Shao-Hua Sun).
Contributed the Covid19CoughAudioClassification task to the Dynamic-SUPERB benchmark by devising and preprocessing a novel dataset for evaluating universal speech models.

Robot Learning Laboratory (RLLab), NTU

type: Undergraduate Researcher period: Nov 2023 - Oct 2024 Taipei, Taiwan Advisor: Prof. Shao-Hua Sun

Developed a multi-agent discussion framework using role-play that boosted LLM creativity scores by 20% over existing methods.
Designed an LLM creativity evaluation benchmark and metrics with 0.7 human-level correlation, validated on over 1,400 text generations.

Wireless and Mobile Networking Laboratory (WMNLab), NTU

type: Undergraduate Researcher period: Feb 2022 - Nov 2023 Taipei, Taiwan Advisor: Prof. Hung-Yu Wei

Utilized RNN, CatBoost, and XGBoost for the prediction and classification of mobile device handovers.
Providing the core dataset for classification models of handovers through live UDP protocol experiments on the MRT.

];