Data Purchasing Challenge Round 2 is going strong To help you make the most of the challenge, we have released a new baseline containing fast heuristic implementations of some simple ideas.
Need some new ideas? Check out these excellent resources created by the community
If these notebooks and explainers help you, don’t forget to the notebook and leave a comment.
Representation Learning: In his notebook, aorhan explains how Representation Learning can be utilised for data label purchase aka Active Learning (AL). You can find the complete explainer over here.
Explainability: How does your model actually learn? Another notebook by aorhan, explain how given a sample you can identify what features contribute to the decision. And how you can improve your model. Read the notebook over here.
Purchase with data anamoly: In this notebook, moto shows how to use anomaly scores to select images to buy. His notebook explains the approach and the result of his experiments.
Labels co-occurence & Image Similarity: In this explainer santiactics performs label co-occurence analysis & Image Similarity using image embeddings. Read the notebook over here.
Sneak peek into the image sample of Round 2: Taking a data visualisation approach sagar_rathod notebook visualises images from different classes and combinations of them. Complete notebook over here.
Additional Resources from Challenge Organisers and Top Participants: We recently hosted a live Town Hall event where the organisers and participants shared their ideas. You can find the recording & resource compilation over here.
Don’t forget to submit your own notebook or resource for the Community Contribution Prize.
Do you have questions about the notebooks & baseline? Drop a comment on this thread to them answered quickly. What approach will you try? Let us know