Data, Data, and more Data

Let’s make use of all the data available!
Merely trading off time populating knowledge base with time annotating training data?
In some ways, No!
- Annotations can be done by anyone: UROPs, our moms, Joe Smoe on the Web, etc.
- Data may not need to be annotated to be useful.
Where is the largest repository of textual data?
The World Wide Web, of course.
Because it’s so big, statistics is our friend.

Let’s make use of all the data available!