Publications

A visually interpretable detection method combines 3-D ECG with a multi-VGG neural network for myocardial infarction identification

Published in Computer Methods and Programs in Biomedicine, 2022

Presents a detection method for myocardial infarction using 3-D ECG and a multi-VGG neural network.

LoGIC: Multi-LoRA Guided Importance Consensus for Multi-Task Pruning in Vision Transformers

Published in Proceedings of the 40th AAAI Conference on Artificial Intelligence (AAAI), 2026

This paper introduces LoGIC, a framework that leverages multi-LoRA guided importance consensus for efficient multi-task pruning in Vision Transformers.

Recommended citation: Chou, Yu-Hong*, Fang, Rui*, Chen, Hsi-Wen, and Chen, Ming-Syan. (2026). "LoGIC: Multi-LoRA Guided Importance Consensus for Multi-Task Pruning in Vision Transformers." Proceedings of the 40th AAAI Conference on Artificial Intelligence (AAAI-26).
Download Paper

Dual Alignment Framework for Few-shot Learning with Inter-Set and Intra-Set Shifts

Published in Annual Conference on Neural Information Processing Systems (NeurIPS), 2025

Introduces DUAL, a framework for few-shot learning that addresses the dual support-query shift challenge, integrating both inter-set and intra-set distribution shifts.

Recommended citation: Jiang, Siyang, Fang, Rui, Chen, Hsi-Wen, Ding, Wei, Xing, Guoliang, and Chen, Ming-Syan. (2025). "Dual Alignment Framework for Few-shot Learning with Inter-Set and Intra-Set Shifts." NeurIPS 2025.
Download Paper

Bi-Level Early Exiting for Generative Document Retrieval

Published in 27th European Conference on Artificial Intelligence (ECAI), 2024

Introduces Bi-Level Early Exiting (BiLEE) for Generative Document Retrieval, significantly enhancing inference speed while maintaining accuracy.

Recommended citation: Fang, Rui, Yeh, Chin-Yuan, Chen, Hsi-Wen, and Chen, Ming-Syan. (2024). "Bi-Level Early Exiting for Generative Document Retrieval." Frontiers in Artificial Intelligence and Applications, Volume 392: ECAI 2024, 4035–4042.
Download Paper

Dual-Triangular QR Decomposition with Global Acceleration and Partially Q-Rotation Skipping

Published in International Conference on Field-Programmable Technology (ICFPT), 2022

Explores an optimized QR decomposition method with global acceleration and rotation skipping.

Recommended citation: Fang, Rui, Jiang, Siyang, Chen, Hsi-Wen, Ding, Wei, and Chen, Ming-Syan. (2022). "Dual-Triangular QR Decomposition with Global Acceleration and Partially Q-Rotation Skipping." 2022 International Conference on Field-Programmable Technology (ICFPT), 1–4.
Download Paper