Ph.D. student (2010-2014) under the supervision of Professor Jürgen Schmidhuber and Dr. Alexander Forster at IDSIA (Dalle Molle Institute for Artificial Intelligence), USI-SUPSI, Manno-Lugano, Switzerland.
M.S. (highest honors) in Computer Science (March 2005), KyungHee University, Korea.
B.E. (highest honors) in Electronics and Telecommunications (March 2003), HCMC University of Technology, HCMC, Vietnam
Research Interests
I am broadly interested in interactive machine learning and artificial curiosity. This includes aspects of online active learning, hierarchical reinforcement learning, deep unsupervised learning, with applications in autonomous developmental robots, computer vision, and natural language processing.
H. Ngo, M. Luciw, A. Forster, J. Schmidhuber. Confidence-Based Progress-Driven Self-Generated Goals for Skill Acquisition in Developmental Robots. Frontiers in Cognitive Science Journal, Oct. 2013. (PDF, Video demos)
H. Ngo, M. Luciw, V. Ngo, J. Schmidhuber. Upper Confidence Weighted Learning for Efficient Exploration in Multiclass Prediction with Binary Feedback. IJCAI 2013, Beijing, China. (PDF)
H. Ngo, M. Luciw, A. Forster, J. Schmidhuber. Learning Skills from Play: Artificial Curiosity on a Katana Robot Arm. IJCNN 2012, Brisbane, Australia. (IM-CLeVeR project page, PDF, Video demos)
H. Ngo, M. Ring, J. Schmidhuber. Compression Progress-based Curiosity Drive for Developmental Learning. ICDL-EpiRob 2011, Frankfurt, Germany. (PDF)
Deep learning of behavioral hierarchy for autonomous visual navigation: On-going work, applying deep RL directly on images from onboard camera to learn robotic skills for navigation. The novelty of our approach lies in the mechanism of self-generating goals, which exploits the environment structure and regularities through a deep hierarchical temporal coherence network.