I presented the first published work from my PhD this week at the Australasian Language Technology Workshop (ALTW) – a paper entitled Classifying Speech Acts using Verbal Response Modes. Happily for me, our paper ended up being judged by the international panel of reviewers as the joint recipient of the Best Paper Award for the conference. Stephen Wan, a fellow PhD student at both Macquarie Uni and CSIRO, was awarded the best student presentation award.

Our paper presented my early work with Robert Dale and Cécile Paris towards automatically classifying intentional structure in text. We approach the problem by automatically classifying speech acts using the Verbal Response Modes (VRM) taxonomy created by Professor William Stiles at Miami University.

As well as reporting on our first classifier of surface speech acts using VRM, the paper also sets the scene for our ongoing research into how we can usefully exploit knowledge of speech acts in email and other forms of online conversation.

In particular, I’m interested in using knowledge of the intentional structure to improve how we currently search conversations and to provide insight into how we might automatically generate summaries of such conversations. I’m also exploring how such structure can provide some automated indication of conversation state. Although our work is still quite preliminary, we got lots of interest and some great feedback and insight from a range of people at ALTW.

If you want to know more about our work, please take a look at the paper. If you have any queries or comments, I’d really appreciate if you would comment on this post or drop me an email (Andrew.Lampert@csiro.au).