Machine Learning Systems
Designing systems that are largely defined by the execution of machine learning workloads present new open problems that straddle the domains of systems, hardware and artificial intelligence. Cyber-physical systems are a prime example of this emerging category. This activity considers the needs of the next generation of machine-learning-centric systems in terms of: design, interfaces and abstractions; parallel, distributed and scalable learning/inference algorithms; hardware co-design for efficiency and high-utilization; and finally, interpretability, security and testing.
DeepTIO: A Deep Thermal−Inertial Odometry with Visual Hallucination
M.R.U. Saputra P.P.B. de Gusmao C.X. Lu Y. Almalioglu S. Rosa C. Chen J. Wahlstrom W. Wang A. Markham and N. Trigoni
In IEEE Robotics and Automation Letters (RAL) + IEEE ICRA. 2020.
Distilling Knowledge From a Deep Pose Regressor Network
Muhamad Risqi U. Saputra Pedro P. B. de Gusmao Yasin Almalioglu Andrew Markham and Niki Trigoni
In IEEE/CVF International Conference on Computer Vision (ICCV). 2019.
Learning Monocular Visual Odometry through Geometry−Aware Curriculum Learning
Muhamad Risqi U. Saputra Pedro P. B. de Gusmao Sen Wang Andrew Markham and Niki Trigoni
In IEEE International Conference on Robotics and Automation (ICRA). 2019.