Class-Balanced Loss Based on Effective Number of Samples

shunk031 / paper-survey

📚 Survey of previous research and related works on machine learning (especially Deep Learning) in Japanese

https://shunk031.github.io/paper-survey/

151 stars 12 forks source link

Class-Balanced Loss Based on Effective Number of Samples #216

Closed shunk031 closed 5 years ago

shunk031 commented 5 years ago

https://arxiv.org/abs/1901.05555

shunk031 commented 5 years ago

実世界のデータセットは long-tail な分布を持つ不均衡データであることが多いが、先行研究ではre-sampling やcost-sensitive learningの観点から解決が図られてきた。こうしたlong-tailな不均衡データに対して、本研究では各クラス数の分布を適切に考慮した class-balanced loss を提案してる。

Class-Balanced Loss Based on Effective Number of Samples - Paper Survey