オーロラは大気構造をエンコードするのか？潜在的レジーム分析と帰属
原題: Does Aurora Encode Atmospheric Structure? Latent Regime Analysis and Attribution
分析結果
- カテゴリ
- 宇宙
- 重要度
- 53
- トレンドスコア
- 12
- 要約
- 機械学習の基盤モデルは、大気のダイナミクスを正確かつ効率的に模倣できるが、その内部は不透明な「ブラックボックス」として機能する。本研究では、これらのモデルがどのように大気構造を内部的に表現しているかを調査し、潜在的なレジーム分析を通じてその帰属を明らかにする。
- キーワード
arXiv:2606.26361v1 Announce Type: new Abstract: ML foundation models are able to emulate atmospheric dynamics accurately and efficiently but operate as opaque ``black boxes''. We investigate the internal representations of the Aurora model using spatially pooled PCA and layer-wise relevance propagation (LRP). We find evidence that Aurora's latent space is primarily organized by seasonal cycles, whereas extreme storm events do not form a linearly separable cluster. LRP indicates that the model attends to features consistent with the 3D vertical structure of the Great Storm of 1987. Perturbation tests show masking relevant regions degrades forecasts $3.31\times$ more than random masking. These findings suggest that Aurora learns meteorological coherence and vertical structure without explicit instruction. arXiv:2606.26361v1 Announce Type: new Abstract: ML foundation models are able to emulate atmospheric dynamics accurately and efficiently but operate as opaque ``black boxes''. We investigate the internal representations of the Aurora model using spatially pooled PCA and layer-wise relevance propagation (LRP). We find evidence that Aurora's latent space is primarily organized by seasonal cycles, whereas extreme storm events do not form a linearly separable cluster. LRP indicates that the model attends to features consistent with the 3D vertical structure of the Great Storm of 1987. Perturbation tests show masking relevant regions degrades forecasts $3.31\times$ more than random masking. These findings suggest that Aurora learns meteorological coherence and vertical structure without explicit instruction.