Asst Prof Harold Soh

Department of Computer Science,
School of Computing, NUS

Research Interests: Human AI/Robot Interaction, Machine Learning and Decision-Making

Contact: 651 67367



Harold Soh is an Assistant Professor in the Dept. of Computer Science at NUS, where he develops artificial intelligence that interacts naturally with people. He envisions teams of humans and intelligent machines cooperating seamlessly to solve the world’s problems: burgeoning healthcare costs, cybersecurity, and climate change. Harold received his PhD from Imperial College London, where his graduate thesis on continuous learning for robotics was nominated for the Eryl Cadwaladr Davis Prize. The incorporation of his algorithms in a smart paediatric wheelchair enabled children with disabilities to independently explore their environment; an empowerment critical for their cognitive and social development. Key to the system’s usability is a learning algorithm that enabled healthcare experts (e.g., occupational therapists) to interactively teach the wheelchair to assist. The system was field-tested at a London hospital and was recognized both by the research community and in the public press: the work was shortlisted for the UK James Dyson Award, featured in MIT Tech Review, and selected for special presentation at ACM/IEEE HRI 2017. Harold has also contributed to improving how robots perceive their environment, specifically via our most primal sense: touch. His work on enabling the iCub robot to incrementally classify objects from tactile sensors was a finalist for the Cognitive Robotics Best Paper award at IEEE IROS 2012. More recently, Harold focuses on enabling intelligent machines to better understand people: he develops machine-learning models of human perception and attributes—e.g., trust, attention, cognition, behaviour—that are integrated into AI learning and decision-making to improve collaborative task outcomes. He is applying his research towards crafting human-centric intelligent assistants for diabetics and their doctors, cybersecurity analysts, and policy-makers.

Current Projects

  • Modelling and Planning with Trust in Human Robot Interaction
  • Bridging Human Teaching and Machine Learning
  • Collaborative AI for with Human Models
  • Artificial-Intelligence Enhanced Next-Generation Robot Skin

Selected Publications

  • Soh, H., Sanner, S., White, M. & Jamieson, G. Deep Sequential Recommendation for Personalized Adaptive User Interfaces. Proc. 22nd Int. Conf. Intell. User Interfaces – ACM IUI ’17 589–593 (2017).
  • Soh, H. Distance-preserving probabilistic embeddings with side information: Variational Bayesian multidimensional scaling Gaussian process. in IJCAI International Joint Conference on Artificial Intelligence, 2011–2017 (AAAI Press, 2016).
  • Soh, H. & Demiris, Y. Learning Assistance by Demonstration: Smart Mobility With Shared Control and Paired Haptic Controllers. J. Human-Robot Interact. 4, 76–100 (2015).
  • Soh, H. & Demiris, Y. Spatio-temporal learning with the online finite and infinite echo-state Gaussian processes. IEEE Trans. Neural Networks Learn. Syst. 26, 522–536 (2015).
  • Soh, H. & Demiris, Y. Incrementally learning objects by touch: Online discriminative and generative models for tactile-based recognition. IEEE Trans. Haptics 7, 512–525 (2014).
  • Soh, H. & Demiris, Y. When and how to help: An iterative probabilistic model for learning assistance by demonstration. in Intelligent Robots and Systems (IROS), 2013 IEEE/RSJ International Conference on 3230–3236 (2013).
  • Soh, H., Su, Y. & Demiris, Y. Online spatio-temporal Gaussian process experts with application to tactile classification. in Intelligent Robots and Systems (IROS), 2012 IEEE/RSJ International Conference on 4489–4496 (2012).

Awards & Honours

  • M.I.T. SMART Postdoctoral Fellowship Award, 2013
  • UK James Dyson Award National Finalist, 2012
  • IEEE/RSJ IROS Cognitive Robotics Best Paper Finalist, 2012
  • Khazanah Global Scholarship (Postgraduate), 2009-2013.
  • Uni. of California Regents Scholarship, 2000-2004

Teaching (2019/2020)

  • CS2040S: Data Structures and Algorithms
  • CS5340: Uncertainty Modelling in AI
Top Top