† denotes co-first authorship.
ALICE: A Multifaceted Evaluation Framework of Large Audio-Language Models’ In-Context Learning Ability
Yen-Ting Piao†, Jay Chiehen Liao†, Wei-Tang Chien†, Toshiki Ogimoto†, Shang-Tse Chen, Yun-Nung Chen, Chun-Yi Lee, Shao-Yuan Lo. arXiv preprint, 2026. [arXiv]
- A three-stage framework that progressively reduces textual guidance to evaluate LALMs’ in-context learning under audio conditioning.
- Built the LALM inference framework; ran experiments; led results analysis.
TabGSL: Graph Structure Learning for Tabular Data Prediction
Jay Chiehen Liao, Cheng-Te Li. IEEE International Conference on Big Data (BigData), 2025. [IEEE Xplore] · [arXiv]
- A contrastive-learning GNN with a transformer-based extractor that jointly learns instance correlations and feature interactions for tabular prediction.
- Outlined a hyperparameter-tuning guide for graph structure learning.
Graph Neural Networks for Tabular Data Learning: A Survey with Taxonomy and Directions
Cheng-Te Li, Yu-Che Tsai, Chih-Yao Chen, Jay Chiehen Liao. ACM Computing Surveys, 2025. [ACM] · [arXiv] · [GitHub]
- Systematic review of GNNs designed and implemented for tabular data learning.
- Presented as tutorials at IEEE ICDE 2023 and The Web Conference (WWW) 2023.
Enhancing PrEP Adherence Through Person-Centred Mobile App Interventions: A Real-World Data and Machine Learning Approach Using UPrEPU Among Gay, Bisexual and Other Men Who Have Sex with Men in Taiwan
Jay Chiehen Liao, Huei-Jiuan Wu, Tsan-Tse Chuang, Tsai-Wei Chen, Carol Strong. Journal of the International AIDS Society, 28(S5), 2025. [paper]
- Leveraged real-world mobile-app logs and user attributes to predict PrEP-protected sexual events among GBMSM, with SHAP-based analysis of adherence factors.
Predictive Modeling for 14-Day Unplanned Hospital Readmission Risk by Using Machine Learning Algorithms
Yu-Tai Lo†, Jay Chiehen Liao†, Mei-Hua Chen, Chia-Ming Chang, Cheng-Te Li. BMC Medical Informatics and Decision Making, 21, 2021. [paper]
- An explainable ML model for 14-day unplanned hospital-readmission risk from EHR data of 24k+ patients, with SHAP-based feature analysis aligned with clinical experience.