LATENT REFERENCES / TAG2
few-shot
Original title: few-shot
This reference note belongs to Tag2 in Latent References, an archive curated by Keigo Yoshida. Its archive region is Analysis. The note preserves its source text and links so that readers can trace the material behind the 3D map.
- Collection
- Tag2
- Archive region
- Analysis
Archived reference note
English translation of the archived note. JP shows the original text. Source links and literal code are retained; the translation does not update or independently verify the source claims.
Few-shot learning addresses a problem setting where different data classes are given during training and testing. Using a few test-data samples, it aims to adapt the model effectively to the test-data classes and accurately predict test data belonging to classes absent during training.
few-shot学習というのは、学習時とテスト時に異なるクラスのデータが与えられるという問題設定で、数枚(a few)のテストデータを使って上手くモデルをテストデータのクラスに適応させて、学習時には無かったクラスに属するテストデータに対して正確に予測を行えるようにすることを目指す手法です。
Source updated 2024-01-01 · Snapshot 2026-10-08