OpenIE Demo

From Knowitall
Revision as of 21:03, 4 June 2012 by Rbart (talk | contribs)

Jump to: navigation, search

Large Projects

  1. Learning algorithm for search result order.
  2. Grouping for queries with multiple free query slots. Might conflict with previous item.
  3. Classify negative and positive extractions.
  4. Create a reduced index by moving some filtering logic from the frontend to the index.
  5. Creating an in-link ratio for removing over-linked entities.
  6. Adding resolver (or other) synonym database
  7. Creating a simple pipeline for adding new data to the demo (e.g. from current news text)
  8. Add OLLIE extractions

Small Projects

  1. Freebase autocompletion for types and entities, preferably not using old hacky method.
  2. Make popovers persistant.
  3. Progress bar on query.
  4. Better logic for filtering out near subset types.
  5. New icon.

Ideas

  1. Show all unique sentences if any sentence contains an extraction instance with confidences > threshold. Presently we don't show any sentences with extractions < threshold.

Goals

  1. Presentable on the smart phone
  2. Queries complete in less than 10 s