After winning the Victor Lesser Dissertation Award 2018, the thesis by Fernando Santos ‘Dynamics of Reputation and the Self‐organization of Cooperation’ won the Best Artificial Intelligence PhD Thesis Award 2017-2018, delivered by APPIA – Portuguese Association for Artificial Intelligence. “I feel extremely happy. It is gratifying to have my work recognised and I am very grateful to the Portuguese Association for Artificial Intelligence (APPIA)”, confesses the INESC-ID researcher, who is also a Postdoctoral Research Fellow at Princeton University.
“My research topic is very important in the area of Artificial Intelligence (AI). The challenge of understanding how to design systems where intelligent agents cooperate autonomously exists for a long time. In my thesis I explored this topic and developed new mathematical and computational models in order to understand the relationship between reputation and cooperation systems”, explains Fernando Santos. “On the other hand, the fact that this work establishes a bridge between AI and other disciplines, weighed heavily in jury’s decision ” said the Técnico alumnus.
This award gives the winner “an immense motivation for my future work”. “I am currently working in a department of Ecology and Evolutionary Biology, where I have the opportunity to interact and work with extraordinary researchers in creating models to understand, for example, epidemiological dynamics or the evolution of behaviours in animal species”, shares the Técnico alumnus. “I believe that the models developed in this area can be adapted and can contribute to understand artificial agents societies, in particular understanding how to design agents that sustain desirable collective dynamics – whether to maintain cooperation or avoid conflict” he adds.
“I believe that with the increasing number of artificial agents making decisions in our daily lives, it is imperative to build models to understand, predict and control behaviours in populations of humans and agents”, he stresses. According to the Técnico alumnus, access to large amount of data and increased computing power will “help to identify the advantages of different social norms in behaviour regulation”. “However, if interaction with robots or virtual agents fundamentally alters human behaviour, the data available today may be insufficient”, he says. Addressing these challenges will require “laboratory experiments to see how people deal with artificial agents, as well as building models to anticipate the effect of these long-term behavioural changes”, says the Técnico alumnus .
“It’s amazing to see the advances and applications of AI day after day and the challenges ahead”, says Fernando Santos, confessing his enthusiasm for working in such a challenging and ever-changing area. “There is a lot of work to be done to understand the fundamental aspects of human intelligence in order to transpose them into a machine”, he adds.