Top

Nov 22nd, 2013, 3:00 pm in Huxley 217/218

What else could you know, given what I assume you know?

In our everyday lives we engage in numerous dialogues as we socialise with people around us. Most times we attempt to persuade others to accept our point of view, which requires that we strategize against them so as to increase our chances of winning. However, as we are unaware of their knowledge we can only rely on our assumptions on what they might know, which are usually based on our previous dialogues with them, i.e. on the arguments they used in those dialogues. So for every individual we come against we can construct a model of their knowledge consisting of all the arguments they ever used against us, and rely on it for strategizing.

But wait a second! Is this all we can do? Could we possibly infer more things about what one might know given what we assume they know? And if yes, how can we do that? In this presentation we explore this possibility through intuitive everyday examples, which appeal to intuition and make conveying the ideas behind our research easy and fun. Need more reasons to attend this talk? How about never losing an argument again?

Christos Hadjinikolis is a senior PhD student in the Department of Informatics at King's College London working on agent dialogue games based on argumentative systems. In his research he focuses on opponent modelling aspects of such games and with strategizing, while he is also interested in the development of Monte-Carlo simulations. Email to me→