Skip to main content

Christopher Iliffe Sprague

Profile picture of Christopher Iliffe Sprague





About me

I am a researcher in artificial intelligence (AI) planning, interested in combining machine learning (ML) and model-based (MB) methods to create autonomous planning systems that are efficient (conserving time, energy, or money), flexible (modular, scalable, easy-to-reprogram), and trustworthy (human-explainable, formally guaranteed to behave as expected).

I have a diverse background in computer science, control theory, and aerospace engineering. Additionally, I have extensive international experience and have worked with both space and maritime robotics. I've also supervised multiple master's students and have designed assignments for master-level AI/robotics courses.

I've published peer-reviewed research ranging across the following selected topics:

  • formally guaranteeing the performance of behaviour trees (BTs), a modular alternative to finite-state machines (FSMs);
  • using deep ML and optimal control theory to learn optimal policies with multiple changing objectives;
  • using deep 3D computer vision architectures to enhance simultaneous localisation and mapping (SLAM) with real point cloud data;
  • using reinforcement learning (RL) for information gathering.

Keywords: machine learning, optimal control, hybrid dynamical systems, graph theory, order theory, robotics, physics-informed learning, behaviour trees.

Software: Python, JAX, Pytorch, Tensorflow, GPyTorch, StableBaselines, SymPy, Mathematica, PyGMO.


Introduction to Robotics (DD2410), assistant | Course web

Machine Learning (DD2421), assistant | Course web