Major CHI milestones

[AbeHandler]

Coding and automatic evaluation

• [ ] Get arora/sbert running on CURC so we can run experiments, experiment w/ hyperparam settings like glove/w2v/embedding size etc. This one has a few substeps:

  • [ ] Fix this one: https://github.com/fortytw0/needfinder/issues/11 .
  • [ ] This would help us with https://github.com/fortytw0/needfinder/issues/6

• [ ] Auto eval

  • [ ] Decide on tuning procedures/hypermaram choices for transformer-based model
  • [ ] implement them
  • [ ] write them down/explain and justify
  • [ ] Undertake => https://github.com/fortytw0/needfinder/issues/14

Human evaluation

• [ ] Based on autoeval, do human eval => https://github.com/fortytw0/needfinder/issues/3

Writing and feedback

• [ ] Write up everything by maybe mid August so we have weeks to edit and get feedback