Harold SOH Soon Hong

Harold SOH Soon Hong is an Associate Professor in the Department of Computer Science at the National University of Singapore (NUS), School of Computing. He serves as Associate Director of the NUS AI Lab and directs the Collaborative, Learning, and Adaptive Robots (CLeAR) Lab. His research focuses on developing trustworthy collaborative robots through advances in machine learning and human-robot interaction.

Education:

• Ph.D. in Artificial Intelligence & Robotics, Imperial College London, UK (2014)
• M.S. in Software Engineering, University of Melbourne, Australia (2005)
• B.ASc. in Computer Science and Economics, University of California, Davis (2004)

Professor Soh's research centers on machine learning and decision-making for trustworthy collaborative robots. His work spans cognitive modeling (particularly human trust) to physical systems (including novel e-skins for tactile perception). He has made significant contributions to human-robot interaction, especially in developing robots that can learn from and collaborate effectively with humans. His research integrates theoretical foundations with practical implementations in real-world robotic systems.

His recent publications demonstrate a strong trend toward diffusion models, tactile sensing technologies, and trustworthy AI systems. The research spans fundamental machine learning advances to practical robotic applications, with particular emphasis on social navigation, human-robot handovers, and out-of-distribution detection. His work increasingly integrates large language models with physical robotic systems, creating new pathways for embodied AI.

Scientific Awards:

• Best Paper Award at IROS 2021 for Extended Tactile Perception
• Best of IEEE Transactions on Affective Computing Award (2021)
• RSS Best Paper Award Finalist (2018)
• HRI Best Paper Award Finalist (2018)
• RSS Early Career Spotlight Award (2023)
• Multiple NUS Annual Teaching Excellence Awards

Professor Soh actively supervises PhD students and mentors undergraduate research projects through FYP and UROP programs. He has developed and taught courses including CS3264 Foundations of Machine Learning and CS5340 Uncertainty Modelling in AI. His teaching philosophy emphasizes developing independent thinkers with strong analytical skills, fundamental computer science knowledge, and clear communication abilities. His students have won multiple Research Achievement Awards and the NUS Outstanding Undergraduate Researcher Prize.

The CLeAR Lab, which he directs, focuses on developing physical and social intelligence for trustworthy robots. Current projects include Octopi (tactile-language models), Arena platform for social navigation, and GRaCE for robotic grasping. The lab has consistently produced high-impact publications at top venues including RSS, ICRA, and NeurIPS.