Doctoral position, Learning data for companion robots
Linköping University, Sweden
Application deadline:
30 January, 2020
Apply
We are currently accepting applications associated with the research project, ‘The ethics and social consequences of AI and caring robots. Learning trust, empathy and accountability’.
The project is led by Ericka Johnson and Katherine Harrison at Tema Genus, Linköping University, Sweden. More information can be found: https://liu.se/en/research/caring-robots
How can robots learn to care when collecting data on relevant humans may be limited for ethical reasons? Or if real data contain bias, on which data should you train your data? Generative machine learning techniques (such as generative adversarial networks (GANs)) offer a solution to problems with “real” data such as scarce availability, labour intensity of data labelling, data biases, or privacy intrusiveness. This project comprises a critical inquiry into the production/collection of data sets used to help companion robots learn, and particularly the possibility of using GANs to assist with this.