September 20, 2008
Processing Creativity at IJWCC08
The 5th International Joint Workshop on Computational Creativity has concluded here in diurnal, delicious Madrid. This was a small and valuable gathering focused on how computers can model the creative process, but embracing a variety of different media and forms: stories, music, movies, visual art, and even interior decorating. Notably, although there was work presented on generating figurative language, no poetry generator was discussed at all. Beyond the work on specific types of creative production, there was some intriguing theoretical work on computational creativity that was also presented. If the proceedings are available online and I find them, I’ll link to them here. In case not, the book has ISBN 978-84-92539-00-0 and was issued as a technical report of the software engineering and AI department of the Universidad Complutense de Madrid. And the two papers I co-authored are up on my site: “Integrating a Plot Generator and an Automatic Narrator to Create and Tell Stories” by me & Rafael Pérez y Pérez, and “Computing Makes the ‘Man’: Programmer Creativity and the Platform Technology of the Atari Video Computer System” by Ian Bogost & me.
Thanks to the work of conference organizers Pablo Gervés (the local), Rafael Pérez y Pérez (Mexico) and Tony Veale (Ireland), the international diversity of the conference was great. Over the three days there were no female presenters, however, a curious aspect of the workshop that has been discussed a bit and should be discussed further.
At the conclusion of the workshop, we had a very useful discussion about how contests could encourage further research and what form a first contest might take. One idea is to develop the sorts of challenges seen in natural language processing, where the precision and recall of parsing can be measured in a consistent way across all competitors, who use a common data set. Since computational creativity does not seem, from my perspective, to be data-driven or to have useful measures of success that are external, applicable across systems, and objective, I’m not yet sold on this idea. More interesting to me would be an art show that could have (as one participant suggested) both juried and audience awards and that would be physically situated at the workshop and open to the public. How to present the generated work so that the process is clear (if we believe process should be considered) is not obvious, although I’ve heard some ideas. I’m glad the workshop included some discussion of how to motivate further interesting research; I think it helped me to understand my own goals as well as being a useful discussion for this field.
September 24th, 2008 at 5:45 pm
Very interesting. The generated stories I found on the Mexica website looked promising. Future developments should be exciting!
October 10th, 2008 at 11:17 am
I was there, and I must say it was a great experience. Also, I was happy to see that there’s a generational change going on between the ‘baldies’ (Geraint A. Wiggins, John Gero, Graeme Ritchie, Simon Colton, maybe even Chris Thornton) and younger people like Will Byrne, Jamie Forth, Mark O. Riedl, and Carlos León, among others. The former seemed to me like these Rock ‘n’ Roll bands that get back together after twenty years of separation because they need the money. What I mean is (to say it politely), that their ideas are not fresh anymore, and they should definitely step aside from the action. Graeme Ritchie’s talk made me feel ashamed (but not as much as he did at the dinner table, I must say). Not even I would have presented something that ridiculous. Or John Gero, who instead of listening to other people’s talks, spent his time surfing the web and buying music on iTunes. (I was sitting behind him.)
The younger people got, on the other hand, a different way of thinking on computational creativity. Mark Riedl came up with a great idea for automatic storytelling (dividing the problem of telling a story into smaller subproblems called vignettes. An analogy could be drawn here between story/integrals, vignette/differentials. At least that’s how I think about them). Will Byrne was the only one who defined computational creativity in terms of problem solving, which, I think, is the way to go in this field. (Most of the baldies are still stuck on the «computationally creative is anything made by a machine that we would deem creative in humans» kind of definition. How is any branch of science going to evolve with such ‘definitions’? I guess that’s the burden of being Maggie Boden’s followers.
One thing that surprised me, was that nobody even mentioned chess programs, or games of any kind for that matter. (No, wait. Nick did. The three times he appeared on stage. By the way, the integration of MEXICA and nn was universally regarded by the part of the audience I could talk to as a match made in heaven.) Anyway, that’s a field of study that could have a bright future. After all, there’s not much money on creative painting programs, but in games, now there’s a lot.