School of Electronics, Electrical Engineering and Computer Science
& ECIT Global Research Institute
Proposed Project Title: Gentleman AI
Principal Supervisor: Deepak Padmanabhan
Modern advances in Artificial Intelligence are centred on the usage of machine learning to learn over massive datasets, so that the AI system becomes smart enough to do meaningful tasks. If you upload a picture, an AI system could automatically suggest tags you may want to use, and sentiment analysis AI can summarize the sentiment of text reviews written by users.
Humans generate a lot of data of undesirable quality; for example, it is common to find racist overtones or offensive language used freely on the web. Can we produce AI systems that can learn over the content without being affected by the undesirable qualities of the data it learns over, so its behaviour would consequently be more gentleman-like? Further, can we develop AI that can alter the content generated by human and de-bias them so they are made more worthy of consumption by humans?
This project attempts to put these questions around ethics and AI at the centrestage, and intends to develop algorithms to address these challenges.
Supervisor Name: Deepak Padmanabhan Tel: +44 (0)28 9097 4874
QUB Address: Computer Science Building Email: email@example.com