Try to figure out why their model works well on extrinsic eval
kudkudak opened this issue · comments
Stanislaw Jastrzebski commented
We talked with Dima:
- Evaluation, if it has negative samples, ultimately sucks, so we need extrinsic evaluation
- Paper also figured it out, it has extrinsic evaluation, and model performs well on it. Weird!
Task is to examine hypotheses around why it happens
Stanislaw Jastrzebski commented
Seems like it doesn't really work on wiki. Next steps:
- Validate that I reproduce good results on ConceptNet tuples
- See closest example in both wiki and ConceptNet