Watch Me Work: Search driven by your own writing.

As we work, we often need information to support our thinking. This often requires turning away from the work, pulling up an engine and then typing in a query. If you are writing, however, the queries that we need are already embedded in the text of the document we are authoring.

This project leverages the text in our own documents to automatically create queries that can then be used to find useful information related to our own work product. Using the terms, the view and the current focus, the document itself will provide the grist for the construction of queries to be applied against multiple engines. The results of these queries can then be organized and presented to end users.

Faculty and Staff Leads

Kris Hammond

Professor of Electrical Engineering and Computer Science

Prior to joining the faculty at Northwestern, Kris founded the University of Chicago’s Artificial Intelligence Laboratory. His research has been primarily focused on artificial intelligence, machine-generated content and context-driven information systems. Kris currently sits on a United Nations policy committee run by the United Nations Institute for Disarmament Research (UNIDIR). He received his PhD from Yale.

Project Details

2018 Fall

Important Questions

How can the structure of a document inform the terms that are important?
How can your position in a document inform the terms that are important?
How can results be presented in a way that does not intrude?
Can automatically generated queries be easily tailored to by end users?
What other aspects of work can be used to provide more context in the formation of queries?
What controls can be put into users’ hands?

Sample Milestones

Week 1: Core use cases and choice of authoring platform (Google Docs?) and search APIs.
Weeks 2-3: Development of initial term extraction and search systems.
Week 4: Integration of baseline extraction and search.
Weeks 5-6: Design of presentation model and integration with extraction/search.
Weeks 7-8: Refinement of extraction model based on position, structure and activity.
Week 9: Exploration of other context models (open files, files in folders, etc.)

Outcome

The development of a system that allows users to find documents related to their current work product without ever crafting explicit queries.