Robotics and Machine Learning

Supervisor: Dr R Gross

Project Description

This project explores a novel artificial intelligence method, called Turing Learning [2], that enables robots to model themselves, each other or their environment without prior information. Turing Learning generalizes GANs - one of the most promising machine learning methods [3], making it applicable to real-time modelling [4].

Turing Learning refers to a family of machine learning algorithms that generate models and discriminators in a competitive setting [2]. In Turing Learning, a training agent, T, provides genuine data samples (the training data), while a model agent, M, provides counterfeit data samples. A discriminator agent, D, labels the data samples as either genuine or counterfeit. Agents M and D are being optimized: D is rewarded for labelling data samples correctly, while M is rewarded for misleading D (to label its data samples as genuine). The underlying idea of Turing Learning is inspired by the Turing test.

The project will be carried out in the Natural Robotics Lab [1].

[2] Turing learning: a metric-free approach to inferring behavior and its application to swarms.
[3] GANs
[4] Generalizing GANs: A Turing Perspective

Prospective applicants should have a degree in a relevant subject (computer science, engineering, machine learning, artificial intelligence, mathematics, physics, robotics etc) and good experience in programming.

Contact Details

Dr R Gross

  • Tel: +44 (0) 114 222 5610

Back to PhD Project List