What would you do with a system that was intelligent enough to analyze a Wikipedia article along with qualitative human judgments of that article ("brilliant prose") and tell you exactly what the humans meant, even when they weren't sure themselves?
At least years Wikimania Erik Zachte and I discussed exactly this possibility. Invariably these discussions lead to the subjective nature of quality, and quickly diverge to [[Zen and the art of Motorcycle Maintenance]]. But my colleagues and I have determined that this problem should be tractable, and have initiated a research program to find out if we are correct.
What we would like to know is your dreams for such a system, what you would like it to do, and what you would do with it. To help kick-start your imaginations, please see the following paper, written by myself, a psychologist, Trevor Pincock, a linguist, and Laura Rassbach, a computer scientist. Although we consider our findings to be preliminary, we would also like to emphasize the phenomenal rate of growth of the field of [[Natural Language Processing]]. We *will* be able to build the system of your dreams. We just need to hear them first.
Please note that this paper is a draft. Please do not cite it.
"Exploring the Feasibility of Automatically Rating Online Article Quality" http://whisper.colorado.edu/RassbachPincockMingus07.pdf
/Brian Mingus en:User:Alterego http://en.wikipedia.org/wiki/User:Alterego
What's the killer scenario for such a system?
One possibility is a system that helps page "gardeners" in weeding out garbage contributions, or modifying them to be more appropriate. But I'm not sure that would be that useful. The track page feature seems good enough to allow page gardeners to follow changes being made and to quicky evaluate them.
Another possibility is something that would asses the page for "unity". For example, many pages contain entries that are written at different levels of formality. Rewriting those pages to make them more unified is hard work. So a tool that would help doing that might be useful.
---- Alain Désilets, National Research Council of Canada Chair, WikiSym 2007
2007 International Symposium on Wikis Wikis at Work in the World: Open, Organic, Participatory Media for the 21st Century
http://www.wikisym.org/ws2007/
________________________________
From: wiki-research-l-bounces@lists.wikimedia.org [mailto:wiki-research-l-bounces@lists.wikimedia.org] On Behalf Of Brian Sent: May 9, 2007 12:38 AM To: Research into Wikimedia content and communities Subject: [Wiki-research-l] Exploring the Feasibility of Automatically RatingOnline Article Quality What would you do with a system that was intelligent enough to analyze a Wikipedia article along with qualitative human judgments of that article ("brilliant prose") and tell you exactly what the humans meant, even when they weren't sure themselves? At least years Wikimania Erik Zachte and I discussed exactly this possibility. Invariably these discussions lead to the subjective nature of quality, and quickly diverge to [[Zen and the art of Motorcycle Maintenance]]. But my colleagues and I have determined that this problem should be tractable, and have initiated a research program to find out if we are correct. What we would like to know is your dreams for such a system, what you would like it to do, and what you would do with it. To help kick-start your imaginations, please see the following paper, written by myself, a psychologist, Trevor Pincock, a linguist, and Laura Rassbach, a computer scientist. Although we consider our findings to be preliminary, we would also like to emphasize the phenomenal rate of growth of the field of [[Natural Language Processing]]. We *will* be able to build the system of your dreams. We just need to hear them first. Please note that this paper is a draft. Please do not cite it. "Exploring the Feasibility of Automatically Rating Online Article Quality" http://whisper.colorado.edu/RassbachPincockMingus07.pdf /Brian Mingus en:User:Alterego http://en.wikipedia.org/wiki/User:Alterego
Interesting paper.
I am building a model of wiki community behavior that talks in terms such as number of reads and writes under various circumstances. My model hypothesizes various mental states that are not so easily measured but might be possible to infer. I'm not far enough along to say much more about this now, except ...
In my dreams, I would find it useful to classify contributions into broad categories based on the nature of the writing such that I could get distinct counts for each. My expectation would be that this would simplify my inference of mental states, on average, even if individual classifications were not reliable.
Best regards. -- Ward
On May 8, 2007, at 9:38 PM, Brian wrote:
What we would like to know is your dreams for such a system, what you would like it to do, and what you would do with it. To help kick-start your imaginations, please see the following paper...
"Exploring the Feasibility of Automatically Rating Online Article Quality" http://whisper.colorado.edu/RassbachPincockMingus07.pdf
wiki-research-l@lists.wikimedia.org