| Abstract |
This is a work in progress talk on the amalgame alignment toolkit
we're developing within the context of the EuropeanaConnect FP7
project. Goal of the talk will be to get feedback and discussion over
questions such as:
- Why do we (or do we not) need yet another alignment toolkit?
- When is automatic alignment a good idea, and when not?
- How to evaluate big (>100k) alignments, or: can we do better than
checking small random samples?
- If domain experts would make better alignments than automatic tools,
what would a tool look like to support them making their own alignments?
I will discuss the questions above (and demo!) how our ideas on what
the answers might be are reflected in the current design of the
amalgame toolkit.
See also http://semanticweb.cs.vu.nl/amalgame/
|