Invited Speakers

Assistant Professor Jiaqing Liu
Ritsumeikan University, Japan 


Biography: Jiaqing Liu received his B.E. degree from Northeastern University, China, in 2016. He then earned both his M.E. (2018) and D.E. (2021) degrees from Ritsumeikan University in Kyoto, Japan. From 2020 to 2021, he was a JSPS Research Fellow for Young Scientists. He subsequently served as a Specially Appointed Assistant Professor at the Department of Intelligent Media, Institute of Scientific and Industrial Research (ISIR), Osaka University, from October 2021 to March 2022. He is currently an Assistant Professor at the College of Information Science and Engineering, Ritsumeikan University. His research interests include computer vision, medical engineering, and deep learning.

 

Speech Title: Multimodal AI for Depression Detection: Advances in Emotion Understanding and Model Interpretability

Abstract: Depression is a prevalent mental health condition that significantly affects quality of life and social functioning. Despite rapid progress in artificial intelligence, current large language models (LLMs) face challenges in integrating multimodal information and offering interpretable emotional insights. In this invited talk, I will present a novel prefix-tuning approach that enables efficient adaptation of emotion-pretrained LLMs for depression detection. Our method models multimodal inputs—such as speech, facial expressions, and text—while tuning only a small subset of parameters. This approach enhances emotional awareness, cross-modal generalization, and model interpretability. It highlights the potential of multimodal AI for mental health applications and lays the groundwork for future emotionally adaptive human-AI interaction systems.
 

 

Associate Professor Yuan Ye
Shenyang University of Technology, China
 

Biography: Yuan Ye currently serves as an Associate Professor in the Software College of Shenyang University of Technology in China. She received her B.E. degree and Master's degree from the Zhejiang University, China, and the University of New South Wales, Australia, in 2010 and 2012, respectively. She obtained her Ph.D. from the Northeastern University, China, in 2022, and had academic exchanges at the Ritsumeikan University, Japan, during 2014-2016. Professor Yuan's primary research interests lie in the fields of image segmentation, 3D reconstruction, and machine learning. She has authored one academic monograph and contributed numerous papers to international journals and conferences, two of which have been published in IEEE Transactions on Image Processing.

 

Speech Title: An Experience-driven Interpretable Multi-task Model for Segmentation and Classification of Small Cell Lung Cancer and Non-small Cell Lung Cancer from CT Images

Abstract: Accurate identification of small and non-small cell lung cancer classes holds significant importance in clinical practice. An experience-driven interpretable multi-task network to concurrently perform segmentation and classification of small cell and non-small cell lung cancer is introduced. The core architecture of this multi-task model is based on StarNet, featuring a shared feature extraction branch and task-specific decoding branches for tumor segmentation and classification. Leveraging clinical knowledge of small cell lung cancer characteristics, such as indistinct edges, tissue invasion, and limited large cavity areas, two auxiliary branches are proposed: edge uncertainty estimation and tumor core area reconstruction. The values from edge uncertainty estimation and reconstruction integrity estimation are utilized in the classification branch to facilitate small cell lung cancer classification. Furthermore, to enhanced interpretability, bottleneck layer features are extracted for comparative learning, and a three-level contrastive loss is proposed to improve the differentiation of disease features. Lastly, an interpretable strategy based on trained feature query matching is presented, providing radiologists with clinical insights and reference images while the model outputs recognition predictions.