Discover the forefront of AI innovation at the SMU-MSRA Joint Workshop, where experts from SMU’s School of Computing and Information Systems (SCIS) and Microsoft Research Asia (MSRA) share cutting-edge advancements and their practical implications. This workshop delves into transformative topics, including equipping AI with diverse human values for ethical decision-making, enhancing robotic models for real-world applications, and leveraging foundation models to revolutionize healthcare diagnostics and outcomes, a great opportunity to gain insights into the future of AI across sectors like finance, science, and high-performance computing, and interact with researchers from both academia and industry. There will be presentations from both SMU and MSRA, followed by open discussions.
Speakers: - Xing Xie, Partner Research Manager, MSRA
Talk title: Value Compass - Equipping AI with Diverse Human Values Abstract: With the rapid development of large language models profoundly transforming our daily lives, concerns about their safety and ethical implications are growing. Ensuring that AI behavior aligns with human values has become a critical topic of widespread interest. In this talk, I will explore some of the key challenges in this field, including how to define human values and evaluate the extent to which models adhere to them. To address these issues, we adopt an interdisciplinary approach, working closely with experts in ethics and sociology to leverage their extensive experience in studying human values. I hope this talk will inspire in-depth discussions and foster greater interdisciplinary collaboration. Bio: Dr. Xing Xie is a partner research manager at Microsoft Research Asia. He received his B.S. and Ph.D. in Computer Science from the University of Science and Technology of China in 1996 and 2001, respectively. Since joining Microsoft Research Asia in July 2001, Dr. Xie has focused on data mining, social computing, and responsible AI. His work has been recognized with several prestigious awards, including the IEEE MDM 2023 Test-of-Time Award, ACM SIGKDD 2022 Test-of-Time Award, ACM SIGKDD China 2021 Test-of-Time Award, ACM SIGSPATIAL 2020 10-Year Impact Award Honorable Mention, and ACM SIGSPATIAL 2019 10-Year Impact Award. He has delivered keynote speeches at notable conferences such as MDM 2019, ASONAM 2017, and W2GIS 2011. Dr. Xie serves on the editorial boards of ACM Transactions on Recommender Systems, ACM Transactions on Social Computing, ACM Transactions on Intelligent Systems and Technology, and CCF Transactions on Pervasive Computing and Interaction. He served as program co-chair of ACM Ubicomp 2011, PCC 2012, UIC 2015, SMP 2017, ACM SIGSPATIAL 2021, ACM SIGSPATIAL 2022, IEEE MDM 2022, and PAKDD 2024. Dr. Xie is a Fellow of the ACM, IEEE, and China Computer Federation.
- Jiang Bian, Senior Principal Research Manager, MSRA
Talk title: Enhancing Generalization Capabilities of Robotic Foundation Models Abstract: We aim to develop advanced methods for improving the generalization capabilities of robotic foundation models, addressing challenges such as converting abstract instructions into actionable inputs, enabling robots to handle diverse tasks, and overcoming the limitations of data diversity and collection costs. We seek to empower robotic agents to learn and adapt to real-world environments by integrating foundational principles derived from multimodal data, such as visual, textual, and spatial cues. This research will enable robotic models to generalize effectively to novel objects and scenarios while preserving robust task execution. Bio: Dr. Jiang Bian serves as a Senior Principal Research Manager at Microsoft Research, where he spearheads the Machine Learning Group and the MSR Asia Industry Innovation Center (MIIC). The mission of his team involves cultivating innovative, disruptive AI technologies, including reinforcement learning and generative AI, and fostering forward-looking insights to accelerate digital transformation across key industries. Specifically, Dr. Bian guides his team in developing groundbreaking machine learning frameworks and algorithms for real-world applications, encompassing finance, logistics, manufacturing, healthcare, energy and sustainability. Dr. Bian holds a bachelor's degree from Peking University in China and earned his Ph.D. in computer science at the Georgia Institute of Technology in the United States.
- Xinxing Xu, Principal Research Manager, MSRA
Talk title: Multimodal Foundation Models for Healthcare Abstract: Foundation models are transforming healthcare by leveraging diverse data, including Electronic Medical Records, medical images (Chest X-rays, CT, MRI, pathology scans), and multiomics data to enhance diagnostics and patient outcomes. This talk covers recent advancements in foundation models, featuring Prov-GigaPath, a whole-slide pathology model pretrained on 1.3 billion image tiles, and BiomedParse, which jointly performs segmentation, detection, and recognition across nine modalities. Additionally, we introduce BenchX, a unified benchmark framework for evaluating Medical Vision-Language Pretraining methods. These innovations pave the way for more accurate, efficient, and scalable healthcare solutions. Bio: Dr Xinxing Xu is a Principal Research Manager at Microsoft Research Asia Singapore. His research interests include machine learning, computer vision, multimodal AI, resource-efficient learning, multimodal medical data analysis and AI for industrial applications, such as digital healthcare and advanced manufacturing. Prior to joining Microsoft, he was a Senior Scientist and the Group Manager of Multimodal AI, Computing & Intelligence Department, and the Innovation Target Area (ITA) Lead, MedTech & HealthTech, at Institute of High Performance Computing (IHPC), The Agency for Science, Technology and Research (A*STAR), Singapore. He also was the Adjunct Assistant Professor at Duke-NUS Medical School of the National University of Singapore (NUS), and the Adjunct Principal Investigator at the Singapore Eye Research Institute (SERI), Singapore National Eye Center (SNEC). He obtained his bachelor's degree in electronic engineering and information science from the University of Science and Technology of China (USTC), Ph.D. in Computer Engineering from Nanyang Technological University (NTU), Singapore.
- Lewen Wang, Researcher, MSRA
Talk title: Large Models in Finance: From Innovation to Real-World Impact Abstract: The wave of large models, driven by the rapid advancements in large language models (LLMs), is now transforming various industries, including finance. This talk will focus on our latest research in applying large models to the financial domain, covering both the development of generative foundation models for financial markets (Large Market Model), which uses order-level generative modeling for realistic market simulations, and practical strategies for leveraging LLMs to empower the finance industry (R&D Agent). Beyond methodological innovations, we will highlight the real-world impact of our research, showing how our work addresses practical challenges and delivers tangible value to the financial sector. Bio: Lewen Wang is a researcher at Microsoft Research Asia. She earned her bachelor's degree in computer science from Northwestern Polytechnical University and obtained her master's degree in computer science from Peking University in 2020. Her research focuses on advancing the application of artificial intelligence in the financial sector, spanning areas such as investment, risk management, fraud detection, and anti-money laundering.
- Xiaotian Gao, Senior Researcher, MSRA
Talk title: Rethink solving partial differential equations in the AI era Abstract: Numerically solving partial differential equations (PDEs) has played a pivotal role in computational sciences and engineering. The rapidly developing AI supercomputers and models inspire us to rethink the way of solving PDEs in the emerging AI era. Firstly, we developed EASIER, a scientific computing framework that makes scientists develop and deploy PDE solvers on AI supercomputers as easily as that of large language models. Leveraging EASIER, scientists are able to achieve extreme performance and scalability without any concerns on low-level details. We also proposed a few neural PDE solvers and physics-informed neural network (PINN) models which learn to solve PDEs without expensive simulated data. They are potential methods to reshape the paradigm of solving PDEs in future. Bio: Dr. Xiaotian Gao is a Senior Researcher at Microsoft Research Asia. Dr. Gao earned his Ph.D. degree in Electrical Engineering from Harbin Institute of Technology in 2018. Prior to joining Microsoft, Dr Gao was a Research Scientist at Intel Labs China. His research focuses on computational sciences and engineering (CSE) and general AI techniques with applications on CSE. Many of his research works have been published in top-tier physics and AI journals and conferences, including Physics of Plasmas, Journal of Applied Physics, Neural Networks, ICML, ICLR, AAAI, AISTATS, etc.
- Rajesh Krishna BALAN, Professor of Computer Science, SCIS
Talk title: Increasing Students’ Wellness through Passively Sensed Wi-Fi Infrastructure Abstract: Professor Rajesh Balan will share his research that uses mobile technologies to passively monitor students’ wellness in a school. Patterns that indicate negative traits such as stress, depression, social exclusion etc. could be identified, and group or community interventions could then be provided to alleviate the issues detected among the students. Bio: Rajesh Balan is a Professor of Computer Science at the Singapore Management University. He has been working in the broad area of mobile systems for over 2 decades. He has served as the program chair of ACM HotMobile, ACM MobiSys, COMSNETS, and IEEE MASS and as the general chair of ACM MobiSys, ACM UbiComp, and COMSNETS. In addition to serving on the program committees of over 50 conferences, he was a member of the UbiComp steering committee and an Associate Editor for IMWUT. He is currently an Associate Editor-in-Chief of IEEE Pervasive Computing and the student outreach director of ACM SIGMOBILE. He is an ACM Distinguished Member and received his Ph.D. in Computer Science from Carnegie Mellon University in 2006.
- Tony TANG, Associate Professor of Computer Science, SCIS
Talk title: Embodied Human-AI Interaction: Lessons from Human Collaboration Abstract: Embodied interaction refers to the way we use our bodies and physical presence to communicate and engage with the world around us. As we engage with AI in embodied contexts (i.e. our everyday lives), I argue that AI agents must not only understand our communicative expressions but also express ideas in a way that aligns with our embodied experience. To achieve fluid and natural interactions, we need to move beyond traditional “chatbot”-style exchanges that rely solely on structured textual turns. In this talk, I will present our recent work exploring how equipping AI agents with a deeper understanding of how humans engage in collaborative communication---where meaning and understanding are co-constructed---can enhance their ability to interpret human intentions and respond in more intuitive and meaningful ways. Bio: Dr Tony Tang is a tenured Associate Professor in the School of Computing and Information Systems at the Singapore Management University. He leads the RICELab group (Rethinking Interaction, Collaboration and Engagement), which uses human-computer interaction research approaches (designing, building and evaluating systems) to explore human-AI interaction in creative, collaborative and embodied activities. Over the past several years, his group has studied by studying collaboration tools, mixed-reality computing, information visualization, digital collaboration environments, gaming, and wellness applications. Previously, he was in the Faculty of Information at the University of Toronto, and prior to this, in the Department of Computer Science at the University of Calgary. He regularly serves on the program committee of ACM; he served as the General Co-Chair of CSCW 2022.
- LEE Min Hun, Assistant Professor of Computer Science, SCIS
Talk title: Enabling Human-AI Collaborative Decision-Making in Health Abstract: The rapid advancements in artificial intelligence (AI) and machine learning (ML) have the potential for transformative applications in healthcare. However, deploying these AI systems effectively in clinical practice remains a significant challenge. In this talk, I will share insights from ongoing research aimed at designing, developing, and evaluating a human-AI collaborative decision support system for physical stroke rehabilitation assessment. Bio: Dr. Min Lee is an Assistant Professor in Computer Science at Singapore Management University. His research investigates how humans and AI can collaborate to improve decision-making and social problems (e.g. health). I explore how we can make AI/ML models more trustworthy, explainable, and interactive. He earned his Ph.D. degree in computer engineering and M.S. degree in machine learning from Carnegie Mellon University.
|
