Exploring the next frontier of AI and machine learning at the Institute of Digital Games

Dr Konstantinos Makantasis, post-doctoral researcher in Artificial Intelligence at the Institute of Digital Games has been awarded a prestigious Marie Sklodowska-Curie Individual Fellowships for his Tamed (Tensor-bAsed Machine learning towards genEral moDels of affect) project. Tamed aims to create new methods and algorithms to realise aspects of general emotional intelligence (that is, the ability to understand and manage your own emotions and those of the people around you), one of the core long-term goals of artificial intelligence and artificial psychology.

Critical factors in the award of the fellowship are the excellence of the researcher, the excellence of the host institution and the novelty of the proposed research. Tamed is unique as it proposes using tensor-based and preference (machine) learning models for capturing general aspects of emotion. The Institute has innovated the use of preference learning, publishing award-winning research as well as collaborating with industry leaders such as Ubisoft and local start-ups such as Modl.ai.

The research uses unconventional approaches and novel methods for discovering entirely new representations of emotional manifestations. By using these representations the aim is to develop models that are capable of generalizing across different contexts, moods, personalities, social settings and modalities of human input data which has been a huge challenge for artificial agents. Something anyone who has used Alexa or Siri can attest to. The methods developed through the project will help investigate the degree to which context-free affect models are possible, consolidating research in Europe and beyond.

“The Institute of Digital Games has innovated with unconventional ways to represent emotion in artificial intelligence, such as preference learning for emotions and emotion representation based on decontextualized information and I’m excited that I’ll be able to focus on tackling understanding and interpreting emotion, one of the main long-term visionary goals in artificial intelligence,” said Dr Makantasis about the new project.

“Historically games have pushed innovation in artificial intelligence, think of Deep Blue, AlphaGo and AlphaStar. Over the 70+ years of this wonderful relationship between AI and games, several algorithms have been developed capable of performing tasks that were thought to be too complex for computers. Tamed has a vision that sounds unthinkable nowadays: making computers detect emotion in general ways and contexts! We’re very proud to host Dr Makantasis’ work at the intersections of AI, computer vision, machine learning and affective computing, that pushes the boundaries of the possible again at the Institute of Digital Games,” explained Prof. Georgios Yannakakis, director of the Institute of Digital Games and supervisor of Dr Makantasis’ Tamed project 

As one of the top-ranked postgraduate programmes in game design, the Institute of Digital Games competes with established research powerhouses such as MIT and frequent research partner NYU. The Institute is also currently working on using AI and games for education in their LearnML Erasmus+ project as well as their H2020 Com-N-Play Science project. 

For more information visit https://youtu.be/BPRlQzHEGk4