Mengchu Zhou
Ph.D. & Dist. Professor, Fellow of IEEE, IFAC, AAAS, CAA and NAI
New Jersey Institute of Technology
MengChu Zhou received his B.S. degree in Control Engineering from Nanjing University of Science and Technology, Nanjing, China in 1983, M.S. degree in Automatic Control from Beijing Institute of Technology, Beijing, China in 1986, and Ph. D. degree in Computer and Systems Engineering from Rensselaer Polytechnic Institute, Troy, NY in 1990. He joined the Department of Electrical and Computer Engineering, New Jersey Institute of Technology in 1990, and is now a Distinguished Professor. His interests are in intelligent automation, robotics, Petri nets, Internet of Things, edge/cloud computing, AI, and big data analytics. He has over 1400 publications including 19 books, over 900 journal papers including over 700 IEEE Transactions papers, 32 patents and 32 book-chapters. He is a recipient of Excellence in Research Prize and Medal from NJIT, Humboldt Research Award for US Senior Scientists from Alexander von Humboldt Foundation, and Franklin V. Taylor Memorial Award and the Norbert Wiener Award from IEEE Systems, Man, and Cybernetics Society, and Edison Patent Award from the Research & Development Council of New Jersey. He is Fellow of IEEE, International Federation of Automatic Control (IFAC), American Association for the Advancement of Science (AAAS), Chinese Association of Automation (CAA) and National Academy of Inventors (NAI).
Existing studies on knowledge distillation focus on teacher-centered methods, in which the teacher network is trained according to its own standards before transferring the learned knowledge to a student one. However, due to differences in network structure between teacher and student ones, the knowledge learned by the former may not be truly desired by the latter. Inspired by human educational wisdom, we propose a Student-Centered Distillation (SCD) method that enables the teacher network to adjust its knowledge transfer according to the student’s true needs. We implement it based on human educational wisdom. The teacher network identifies and learns the knowledge desired by the student network on the validation set, and then transfers it to the latter through the training set. To address the problems of current deficiency knowledge, hard sample learning and knowledge forgetting faced by a student network in the learning process, we introduce and improve Proportional-Integral-Derivative (PID) algorithms from the field of control systems to make them effective in identifying the current knowledge required by the student network. Furthermore, we propose a curriculum learning-based fuzzy strategy and apply it to the proposed PID control algorithm, such that the student network can actively pay attention to the learning of challenging samples. Experimental results show that SCD outperforms existing teacher-centered ones in image processing tasks.
Mengchu Zhou
Ph.D. & Dist. Professor, Fellow of IEEE, IFAC, AAAS, CAA and NAI
New Jersey Institute of Technology
MengChu Zhou received his B.S. degree in Control Engineering from Nanjing University of Science and Technology, Nanjing, China in 1983, M.S. degree in Automatic Control from Beijing Institute of Technology, Beijing, China in 1986, and Ph. D. degree in Computer and Systems Engineering from Rensselaer Polytechnic Institute, Troy, NY in 1990. He joined the Department of Electrical and Computer Engineering, New Jersey Institute of Technology in 1990, and is now a Distinguished Professor. His interests are in intelligent automation, robotics, Petri nets, Internet of Things, edge/cloud computing, AI, and big data analytics. He has over 1400 publications including 19 books, over 900 journal papers including over 700 IEEE Transactions papers, 32 patents and 32 book-chapters. He is a recipient of Excellence in Research Prize and Medal from NJIT, Humboldt Research Award for US Senior Scientists from Alexander von Humboldt Foundation, and Franklin V. Taylor Memorial Award and the Norbert Wiener Award from IEEE Systems, Man, and Cybernetics Society, and Edison Patent Award from the Research & Development Council of New Jersey. He is Fellow of IEEE, International Federation of Automatic Control (IFAC), American Association for the Advancement of Science (AAAS), Chinese Association of Automation (CAA) and National Academy of Inventors (NAI).
Existing studies on knowledge distillation focus on teacher-centered methods, in which the teacher network is trained according to its own standards before transferring the learned knowledge to a student one. However, due to differences in network structure between teacher and student ones, the knowledge learned by the former may not be truly desired by the latter. Inspired by human educational wisdom, we propose a Student-Centered Distillation (SCD) method that enables the teacher network to adjust its knowledge transfer according to the student’s true needs. We implement it based on human educational wisdom. The teacher network identifies and learns the knowledge desired by the student network on the validation set, and then transfers it to the latter through the training set. To address the problems of current deficiency knowledge, hard sample learning and knowledge forgetting faced by a student network in the learning process, we introduce and improve Proportional-Integral-Derivative (PID) algorithms from the field of control systems to make them effective in identifying the current knowledge required by the student network. Furthermore, we propose a curriculum learning-based fuzzy strategy and apply it to the proposed PID control algorithm, such that the student network can actively pay attention to the learning of challenging samples. Experimental results show that SCD outperforms existing teacher-centered ones in image processing tasks.
Hiroshi Ishiguro
Professor of Department of Systems Innovation, Osaka University Visiting director of ATR Hiroshi Ishiguro Laboratories
Hiroshi Ishiguro received a Ph. D. from Osaka University, Japan in 1991. He is currently Professor of Department of Systems Innovation at Osaka University, Visiting Director of Hiroshi Ishiguro Laboratories at the Advanced Telecommunications Research Institute (ATR), Project Manager of MOONSHOT R&D Project, Thematic Project Producer of EXPO 2025 Osaka, Kansai, Japan, and CEO of AVITA, Inc. His research interests are interactive robotics, avatar, and android science. Geminoid is an avatar android that is a copy of himself. In 2011, he won the Osaka Cultural Award. In 2015, he received the Prize for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology. He was also awarded the Sheikh Mohammed Bin Rashid Al Maktoum Knowledge Award in Dubai in 2015. Tateisi Award in 2020, and honorary doctorate of Aarhus university in 2021.
In this lecture, the speaker will introduce the technologies of autonomous robots and avatars (teleoperated robots and CG characters) that he has developed, and discuss how the avatar technology will change the world.
It is important to enable people with various backgrounds and values, such as those who need to care for or raise children or the elderly, to participate in diverse activities according to their own lifestyles, and to realize a society in which people are free from the constraints of body, brain, space, and time. Avatar is the realization of such a society.
The speaker is engaged in research, development, and business with the aim of enabling anyone to freely and remotely control multiple avatars and participate in diverse work, educational, medical, and everyday social activities without having to go to the field.
Danielle Belgrave
VP of AI and Machine Learning, GSK
Danielle Belgrave is a VP of AI and Machine Learning at GSK where she leads the AI/ML Clinical Development team. Her career in machine learning for healthcare spans almost 20 years, leading and conducting machine learning research across academia and industry focused on scientific discovery and personalising interventions in health. Before joining GSK, Danielle held research positions at Google DeepMind, Microsoft Research, and Imperial College London. She earned a BSc in Business Mathematics and Statistics from London School of Economics and Political Science, an MSc in Statistics from University College London, and completed both her PhD and postdoctoral research in Machine Learning for Healthcare at University of Manchester. Her contributions have been recognized with numerous prestigious honors, including the Barry Kay Award from the British Society for Allergy and Clinical Immunology, the highly competitive Medical Research Council Award in Biostatistics, and the GSK Exceptional Scientist Award. Danielle is also a distinguished leader in the AI research community. She served as General Chair of NeurIPS 2025 and currently serves on the NeurIPS Board. Previously, she was Program Chair for NeurIPS 2022 and NeurIPS 2024. In addition, she serves on several academic advisory boards, including the Informed AI Advisory Board at University of Bristol and the Foundational AI Centre for Doctoral Training at University College London.
Artificial intelligence is rapidly transforming biomedical research, with foundation models emerging across pathology, genomics, single-cell biology, electronic health records, and scientific literature. Yet despite remarkable advances in individual domains, their impact on drug development has been more limited than many anticipated. The challenge is no longer building capable models, but connecting them into systems that can support scientific and development decisions.
In this keynote, I will explore how foundation models, disease-specific large language models, biomarkers, multimodal patient representations, and agentic AI systems can be integrated into a coherent scientific intelligence layer for drug development. Drawing on real-world experience deploying AI across pathology, translational medicine, and clinical development, I will discuss both the opportunities and the practical challenges of moving from isolated AI capabilities to end-to-end scientific systems.
I will show how agentic scientific systems—AI systems capable of reasoning, planning, and acting across multiple sources of biological and clinical information—are emerging as the connective tissue between foundation models and decision-making. These systems have the potential to accelerate target identification, biomarker discovery, patient stratification, and clinical trial design while enabling scientists to interact with increasingly complex bodies of knowledge and data.
At the same time, significant challenges remain. Biological systems operate across multiple scales, model performance does not always translate into clinical impact, and organizational barriers often limit the adoption of AI innovations. Addressing these challenges requires advances not only in machine learning, but also in scientific evaluation, human-AI collaboration, and the integration of AI into the drug development process.
I will conclude by outlining a path toward integrated AI systems capable of learning unified representations of biological and disease state across modalities and scales, toward what may eventually become Foundation Models of Health. Realizing this vision will require as much scientific and organizational innovation as technical progress.
Hiroshi Ishiguro
Professor of Department of Systems Innovation, Osaka University Visiting director of ATR Hiroshi Ishiguro Laboratories
Hiroshi Ishiguro received a Ph. D. from Osaka University, Japan in 1991. He is currently Professor of Department of Systems Innovation at Osaka University, Visiting Director of Hiroshi Ishiguro Laboratories at the Advanced Telecommunications Research Institute (ATR), Project Manager of MOONSHOT R&D Project, Thematic Project Producer of EXPO 2025 Osaka, Kansai, Japan, and CEO of AVITA, Inc. His research interests are interactive robotics, avatar, and android science. Geminoid is an avatar android that is a copy of himself. In 2011, he won the Osaka Cultural Award. In 2015, he received the Prize for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology. He was also awarded the Sheikh Mohammed Bin Rashid Al Maktoum Knowledge Award in Dubai in 2015. Tateisi Award in 2020, and honorary doctorate of Aarhus university in 2021.
In this lecture, the speaker will introduce the technologies of autonomous robots and avatars (teleoperated robots and CG characters) that he has developed, and discuss how the avatar technology will change the world.
It is important to enable people with various backgrounds and values, such as those who need to care for or raise children or the elderly, to participate in diverse activities according to their own lifestyles, and to realize a society in which people are free from the constraints of body, brain, space, and time. Avatar is the realization of such a society.
The speaker is engaged in research, development, and business with the aim of enabling anyone to freely and remotely control multiple avatars and participate in diverse work, educational, medical, and everyday social activities without having to go to the field.
Danielle Belgrave
Ph.D. & Dist. Professor, Fellow of IEEE, IFAC, AAAS, CAA and NAI
New Jersey Institute of Technology
Danielle Belgrave is a VP of AI and Machine Learning at GSK where she leads the AI/ML Clinical Development team. Her career in machine learning for healthcare spans almost 20 years, leading and conducting machine learning research across academia and industry focused on scientific discovery and personalising interventions in health. Before joining GSK, Danielle held research positions at Google DeepMind, Microsoft Research, and Imperial College London. She earned a BSc in Business Mathematics and Statistics from London School of Economics and Political Science, an MSc in Statistics from University College London, and completed both her PhD and postdoctoral research in Machine Learning for Healthcare at University of Manchester. Her contributions have been recognized with numerous prestigious honors, including the Barry Kay Award from the British Society for Allergy and Clinical Immunology, the highly competitive Medical Research Council Award in Biostatistics, and the GSK Exceptional Scientist Award. Danielle is also a distinguished leader in the AI research community. She served as General Chair of NeurIPS 2025 and currently serves on the NeurIPS Board. Previously, she was Program Chair for NeurIPS 2022 and NeurIPS 2024. In addition, she serves on several academic advisory boards, including the Informed AI Advisory Board at University of Bristol and the Foundational AI Centre for Doctoral Training at University College London.
Artificial intelligence is rapidly transforming biomedical research, with foundation models emerging across pathology, genomics, single-cell biology, electronic health records, and scientific literature. Yet despite remarkable advances in individual domains, their impact on drug development has been more limited than many anticipated. The challenge is no longer building capable models, but connecting them into systems that can support scientific and development decisions.
In this keynote, I will explore how foundation models, disease-specific large language models, biomarkers, multimodal patient representations, and agentic AI systems can be integrated into a coherent scientific intelligence layer for drug development. Drawing on real-world experience deploying AI across pathology, translational medicine, and clinical development, I will discuss both the opportunities and the practical challenges of moving from isolated AI capabilities to end-to-end scientific systems.
I will show how agentic scientific systems—AI systems capable of reasoning, planning, and acting across multiple sources of biological and clinical information—are emerging as the connective tissue between foundation models and decision-making. These systems have the potential to accelerate target identification, biomarker discovery, patient stratification, and clinical trial design while enabling scientists to interact with increasingly complex bodies of knowledge and data.
At the same time, significant challenges remain. Biological systems operate across multiple scales, model performance does not always translate into clinical impact, and organizational barriers often limit the adoption of AI innovations. Addressing these challenges requires advances not only in machine learning, but also in scientific evaluation, human-AI collaboration, and the integration of AI into the drug development process.
I will conclude by outlining a path toward integrated AI systems capable of learning unified representations of biological and disease state across modalities and scales, toward what may eventually become Foundation Models of Health. Realizing this vision will require as much scientific and organizational innovation as technical progress.
@ Copyright - IEEE SMC 2025
Terms & Conditions lEEE Privacy Policy
@ Copyright - IEEE SMC 2026
