Keynote Speaker
Prof. Toshihiko Yamasaki
The University of Tokyo, Japan
Toshihiko Yamasaki received the B.S. degree in electronic engineering, the M.S. degree in information and communication engineering, and the Ph.D. degree from The University of Tokyo. He is currently a Professor at Department of Information and Communication Engineering, Graduate School of Information Science and Technology, The University of Tokyo. He was a JSPS Fellow for Research Abroad and a visiting scientist at Cornell University from Feb. 2011 to Feb. 2013. His current research interests include attractiveness computing, multimedia, computer vision, pattern recognition, and so on. Dr. Yamasaki is a member of IEEE, ACM, AAAI, IEICE, ITE, IPSJ, and JSAI.
Invited Speakers
Invited Speaker I
Assoc. Prof. Tatsuhito Hasegawa
University of Fukui, Japan
Tatsuhito Hasegawa received his Ph.D. degree in Engineering from
Kanazawa University, Japan, in 2015. He worked at Fujitsu Hokuriku
Systems Ltd. and Tokyo Healthcare University before joining the
University of Fukui as a Lecturer in 2017. He has been an Associate
Professor at the University of Fukui since 2020.
His research interests include deep learning, human activity
recognition, mobile and ubiquitous computing, and intelligent
learning support systems. He is currently a researcher in the PRESTO
program of the Japan Science and Technology Agency (JST). He is a
member of IEEE, the Information Processing Society of Japan (IPSJ),
the Japanese Society for Artificial Intelligence (JSAI), and the
Japanese Neural Network Society (JNNS).
Invited Speaker II
Assoc. Prof. Krung Sinapiromsaran
Chulalongkorn University, Thailand
Krung Sinapiromsaran is an Associate Professor in the Department of
Mathematics and Computer Science at Chulalongkorn University,
Bangkok, Thailand. He holds a Ph.D. in Computer Sciences from the
University of Wisconsin-Madison (2000) and has built a distinguished
research career spanning data science, machine learning,
optimization, and knowledge discovery. His work has garnered over
2,700 citations (h-index: 14), with signature contributions
including Safe-Level-SMOTE and DBSMOTE — foundational techniques in
the class imbalance literature that together account for more than
2,000 citations.
His current research focuses on deep generative oversampling for
imbalanced learning, an area he has advanced through a progression
of novel frameworks. DeepCLSMOTE (PeerJ Computer Science, 2026)
introduced class-centroid-optimized autoencoders to improve
latent-space structure for minority class synthesis. Building on
this, CoSMO (under revision) — a Contrastive Synthetic Minority
Oversampling framework — integrates self-contrastive centroid loss
with latent-space SMOTE to address severe class imbalance in
dermatological image classification, achieving a macro F1-score of
0.81 with Vision Mamba on the DermaMNIST benchmark across five deep
learning architectures. His latest work, DeepPolyCL-SMOTE (under
revision), further advances the framework by incorporating
class-balanced weighted supervised contrastive learning and
polynomial-fit SMOTE for topology-aware interpolation, outperforming
BAGAN, GAMO, DeepSMOTE, and DeepCLSMOTE on multiple benchmark
datasets.
In his invited talk at MLHMI 2027, Prof. Sinapiromsaran will present
the unified research thread connecting these three frameworks — from
class-centroid latent structuring to contrastive representation
learning and manifold-aligned oversampling — offering both
theoretical insights and practical guidelines for building reliable
deep learning systems under severe class imbalance in real-world
applications.