Simon.BuckinghamShum.net

I’m just back from COMMA 2010, the 3rd International Conference on Computational Models of Argument. This is an interesting mix of formal AI logic, argumentation theory, and social/semantic web. The logic guys examine how argumentation can be modelled in order to enable the evaluation of arguments based on the relationships between contributions to a debate. The argumentation theory people seek to translate what we know about everyday arguments into schemes that a computer can process, while the social/semantic web people try to figure out how to deliver all this in a palatable, scaleable, interoperable form.

I presented the latest work with Maarten Sierhuis at NASA (now PARC) in which we are introducing software agents, from his powerful Brahms system, into the Compendium hypermedia application that he originally created with Al Selvin while at NYNEX, before it moved to the Open University to continue development, as our research agendas converged. Building on our previous collaboration (Project/Paper), we report two new projects, conducted with Jack Park and Matt Brown:

Buckingham Shum, S.; Sierhuis, M.; Park, J. and Brown, M. (2010). Software Agents in Support of Human Argument Mapping. In: Computational Models of Argument: Proceedings of COMMA 2010, (Eds.) Pietro Baroni, Federico Cerutti, Massimiliano Giacomin, Guillermo Simari. Frontiers in Artificial Intelligence and Applications Series,  Volume 216. Amsterdam: IOS Press (ISSN 0922-6389). http://oro.open.ac.uk/22344

My work sits at the applications end of the continuum in the COMMA design space which I sketch in the opening slides (see the ODET-2010 workshop for others in this space), with this paper exploring how more formal agent-based work might enhance otherwise passive argument mapping tools.

For other work in the socio-semantic web space see Dundee’s ARG group on pipelining argumentation services, Liverpool’s team analysing Amazon review debates, and DERI’s Jodie Schneider envisioning Web 3.0 argumentation. It was a pleasure to catch up with Doug Walton, who gave a succinct overview of the seven core types of dialogue that he and others have characterised, and the role that burden of proof plays (or not) in each case.

I look forward to ongoing conversations with this community.