Data Augmentation Should Follow the Classification Task: A Categorical View on Quotient Mismatch in Neural Classification

青木 邦弘

講演情報

[E] 口頭発表

セッション記号 M (領域外・複数領域) » M-TT 計測技術・研究手法

[M-TT45] Artificial Intelligence in Earth and Environmental sciences

2026年5月24日(日) 15:30 〜 17:00 展示場特設会場 (8) (幕張メッセ国際展示場 7・8ホール)

座長:Kondrashov A Dmitri(University of California Los Angeles)、Richter Ingo(JAMSTEC Japan Agency for Marine-Earth Science and Technology)、Krinitskiy Mikhail(Shirshov Institute of Oceanology, Russian Academy of Sciences)

16:30 〜 16:45

[MTT45-11] Data Augmentation Should Follow the Classification Task:
A Categorical View on Quotient Mismatch in Neural Classification

*青木邦弘¹ (1. 気象庁　気象研究所)

キーワード：深層学習、データ拡張、分類、データ駆動科学

« このセッションの前の講演
このセッションの次の講演 »

▲

講演情報

[M-TT45] Artificial Intelligence in Earth and Environmental sciences

[MTT45-11] Data Augmentation Should Follow the Classification Task: A Categorical View on Quotient Mismatch in Neural Classification

[MTT45-11] Data Augmentation Should Follow the Classification Task:
A Categorical View on Quotient Mismatch in Neural Classification