Implement architecture that jointly learns translation and alignment without supervison
Open
Nobody has claimed this yet.
enhancement
needs to be revisited
- Dominant language
- Python
- Stars
- 39
- Forks
- 7
- Avg merge
- 1d 9h
- Merged PRs (30d)
- 5
Description
The current "Alignment Enhanced" architecture that is implemented in silnlp. This architecture requires already aligned data to train on. It would be better if there was an architecture that could learn to align in an unsupervised manner. See https://arxiv.org/abs/2004.14675.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading the current "Alignment Enhanced" architecture in silnlp and the paper at https://arxiv.org/abs/2004.14675. Compare their requirements for aligned data and unsupervised alignment, then define what evidence would show that the new architecture jointly learns translation and alignment without supervision.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100