More control over loss functions
Open
Nobody has claimed this yet.
- Dominant language
- Java
- Stars
- 352
- Forks
- 38
- PR merge metrics
- No merged PRs in 30d
Description
Reading about multilabel problems because of inbalanced label-distribution in training data:
https://arxiv.org/abs/2109.04712
A label-based loss-function re-weighting is proposed. It seems that at least applying a weights-array is possible with dl4j? How could I do that with dagli?
Contributor guide
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 with the linked paper and DAGLi's existing model-training and loss-function entry points; compare the requested label-based weighting with the current API and the referenced DL4J behavior. Done means establishing how a weights array could be supplied for multilabel training and covering the resulting behavior with the project's existing tests.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100