Grants:TPS/User:とある白い猫/Presenting at PAN Lab of CLEF 2011
Jump to navigation Jump to search
|This Wikimedia Participation Support request was funded in the fiscal year 2011-12. A report is available.|
- Legal name of chapter or nonchapter group
- Grant contact name
- Grant contact username or email
- Grant contact title (position)
- Project lead name
- Project lead username or email
- Project lead title (position), if any
- Researcher, Master of Artificial Intelligence student.
- Full project name
- Amount requested (in USD)
- about 939 EUR which is about 1278 USD
- Provisional target start date
- 19 September 2011
- Provisional completion date
- 22 September 2011
- Travel (train ticket): 90 EUR which is 122 USD
- Accommodations (venue + hotel): 500 EUR which is 680 USD
- Conference registration: 300 EUR which is 408 USD per official rates. Clef2011 is hosting the PAN 2011 Lab
- Poster cost 49 EUR which is 67 USD
- The projects scope is intended for the presentation and observation of work on Automated Wikipedia Vandalism Detection at CLEF 2011's PAN 2011 Lab workshop.
- Own work (Vandal Sense) will be presented at the same workshop.
- The intended goal of this project is to exchange ideas on challenges of automated vandalism detection. It is my belief that an experienced wikimedian and engineer such as myself would contribute to the cumulative discussion. Reports of works presented, including my own, will be published with an ISBN number.
- My past work on automated detection involved IRC bots that detected simple (but common) vandalism based on fixed values. This new approach (VandalSense) relies on machine learning through supervised learning. Work will be resumed on VandalSense based on the observations as I intend to extend the work I have done so far on VandalSense as my Master Thesis on Artificial Intelligence.
- None is required.
Fit to strategy
- Vandalism as well known is among the most important problems Wikimedia projects are facing. The project intends to explore possibilities of dealing with vandalism through the use of machine learning through supervised learning.
- The project would generate a report for the workshop on the discussions which would be of use to people interested in research for this purpose. This report will bring more attention to this field of research.
Measures of success
- This project will be considered a success if enough information is accumulated for further work on VandalSense and if there is an increased awareness of challenges that need to be tackled for more feasible solutions to automated vandalism detection.