The Research team in collaboration with the ML-Platform team are creating a new service to help Wikidata patrollers to detect revisions that might be reverted.
Requirments:
- Model should be able to run in Lift Wing
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | diego | T333892 Develop a new generation of ML models for Wikidata | |||
| Open | diego | T328813 Develop a ML-based service to detect vandalism on Wikidata | |||
| Open | diego | T341820 Evaluate and improve the Revert Risk model for Wikidata. | |||
| Open | None | T343419 Move Wikidata tools to Lift Wing | |||
| Resolved | noarave | T343731 Update API calls from ORES to Lift Wing | |||
| Open | MunizaA | T344016 Improvements to Annotool | |||
| Resolved | BUG REPORT | MunizaA | T343973 Fix relative links for Qids in Annotool | ||
| Resolved | MunizaA | T344152 Allow users to edit annotations in Annotool | |||
| Resolved | MunizaA | T348666 Add randomization to the revision order showed in Annotool | |||
| Open | None | T349739 [Annotool] Include additional information on private project exports |
Event Timeline
Comment Actions
Updates
- We are working on manually evaluating reverts to identify the right data to train the model.
Comment Actions
Update
- Still working on the data evaluation. Currently I'm studying the use of tags and user groups and their relation with reverts.
Comment Actions
Update
- Currently I'm working on featuring engineering. The current model has around 72% accuracy on balanced data.
Comment Actions
Update
- New features had slightly improved the accuracy (now is 75%), I'm still working on improving the model.
Comment Actions
Update
- I'm testing a Deep Learning approach, to see if offers relevant advantages over the current XGBOOST model.