66. Machinic life self-learning and its role on personalised health
Dr. Jose Sousa, PhD – Personal Data Science Team Leader, Sano Science
Abstract
Today machines impact our day life, which is characterised by constantly changing information, by making it simpler, quicker and more comfortable. Imagine what would it be like not to have GPS and rely only on information about directions provided by other humans. Artificial Intelligence (AI) in the beginning (1950s) was grounded in cybernetics and automata, where a machinic life, unlike framed mechanical forms, has a capacity to adapt to concept shifts and to respond dynamically to changing environment. However, AI in its early days would move in the direction of understanding computers as a processor of symbols whereas mechanic life (ALife) (cybernetic systems) would focus on the role and function of computation. The real world is non-stationary, where the rules and relationships of today may change tomorrow. To be able to grasp these changes, and knowing how to tell the difference, remains a big challenge for AI. Humans have effectively managed to do this by having a two-step approach as described in the work of Daniel Kahneman and Antonio Damasio. Firstly, the problem is framed and previous experiences come into play to support the decision (System 1), and secondly, the cognition (System 2) is called into action only if the homeostasis of the decision making is compromised and new knowledge is needed.
Here we introduce a new way how these concepts can be integrated in a machinic life to enhance the transition in personalised healthcare to move from treatment to prevention.
About the author
Dr. Jose Sousa, PhD, was the Manager of the Advanced Informatics Core Technology Unit in the Faculty of Medicine, Health and Life Sciences (FHMLS) at QUB. He obtained his PhD under Prof. Ricardo Machado (University of Minho, Portugal) and Prof. Jose Mendes (University of Aveiro, Portugal) supervision at the University of Minho, Portugal on developing complex network models to study software usage alignment with the project requirements. Previously and during his PhD he worked as Information Systems Director at I3S, a research institution of the University of Porto (i3.sup.pt) where he deployed and managed all the IT infrastructure as well as deploying and developing software to support management and research operations. He was a collaborator of HDRUK – Swansea/QUB substantial site as AI researcher and actively working with QUB Centre for Public Health as part of the QUB support to Northern Ireland Public Health response to pandemic where he has developed self-learning AI on publicly available and self-reporting data. He also has worked on genetic alignment modelling and on mining socio-technical systems.