My main goal is to develop fully autonomous agents that have a human-like ability to understand their environment and act lawfully/ethically while perusing their mission. This requires an agent to understand the law, regulations and general behavioural policies related to the environment it is in. It is pre-requisite that the agent (e.g., physical robot) has the ability to adhere to the restrictions by design. A well-mannered, socially acceptable agent (from a human perspective) should be able to integrate knowledge of the law, regulations and general behavioral policies when planning for task execution. This attribute will be critical in minimising possible legal penalties incurred due to an agent’s action, especially once physical robots become a significant part of society. It will also ensure that agent behaviour does not cause social discomfort or other reasonable concerns.

Interests
  • Reinforcement Learning
  • Robotics
  • Computer Vision
  • Sequence Modelling