[FEA] Compute ground truth scalably on massive-scale datasets
Nobody has claimed this yet.
- Dominant language
- Cuda
- Stars
- 854
- Forks
- 236
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 62
Description
As data sizes continue to grow (exponentially), we're seeing an increased need to be able to compute ground truth efficiently at massive-scale (for eg. 10B, 100B) without having to do an exhaustive brute-force search to find it.
This is a placeholder issue to track work associated with computing ground truth at scale.
Initially, we aim to evaluate breaking a dataset up into clusters (like an IVF) and using only the closest clusters to generate the ground truth.
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
This is a placeholder for scalable ground-truth computation on 10B- to 100B-scale datasets. Start by identifying the cuVS ground-truth entry points and how datasets are currently searched; the issue does not name files or tests, and completion criteria for the clustering approach still need to be defined.
Written by the indexing model from the issue text.
Assessment
- Domain
- data, machine-learning, search
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100