WinterMilestone 3 Team Enigma DarkHorseIdea: Generate tips based on accepted tasks

From crowdresearch
Revision as of 19:18, 31 January 2016 by Dineshdhakal (Talk | contribs) (Dark Horse idea)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Dark Horse idea

While each requester expects their tasks to be done in a certain way, there are widely accepted standards for different types of tasks. We observed a similar pattern in discussions on various forums. So we propose a solution to have a similarity index for various tasks.

  • The aim here is to reduce the amount of time and effort put in by the requester to get better results from the workers
  • Generate a similarity index for tasks from multiple submitters


  • Now based on the above similarity index, when serving a task to a worker, generate a tip or a a skeleton of the expected format of submission.
  • The tips can be based on structural or contextual pattern recognition and learning on the submissions that have already been accepted for similar tasks.
  • Naturally the algorithm should take into account the submissions that were rejected.
  • Eg:- A task that needs an article for product marketing. The simplest tip that could be generated is the approximate number of words expected.

  • The similarity index will be higher when the same requester posts similar tasks in future
  • A more customized form of tip can be provided for the workers when such tasks come up
  • The same model can be further utilized to Automatically Return the tasks to the users when the submission has a higher inclination to previously rejected submissions



Archana Kumari - @dhankie

Dinesh Dhakal - @dineshd