post
Homography
· ☕ 1 min read
ホモグラフィ行列の推定 これで少なくとも必要条件は求まる $s \begin{bmatrix} x^{’} \ y^{’} \ 1 \end{bmatrix} = H \begin{bmatrix} x \ y \ 1 \end{bmatrix} = \begin{bmatrix} h_{11} & h_{12} & h_{13} \ h_{21} & h_{22} & h_{23} \ h_{31} & h_{32} & h_{33} \end{bmatrix} \begin{bmatrix} x \ y \ 1 \end{bmatrix}$ ...
活性拡散
· ☕ 1 min read
知識グラフ / GNN https://www.sciencedirect.com/topics/computer-science/spreading-activation#:~:text=Spreading%20activation%20is%20the%20name,on%20the%20associations%20among%20chunks. ...
Mercer's theorem
· ☕ 1 min read
積分作用素を定義 https://en.wikipedia.org/wiki/Mercer's_theorem https://yagami12.hatenablog.com/entry/2018/12/02/192137#Mercerの定理 ...
Better plain ViT baselines for ImageNet-1k
· ☕ 1 min read
The main differences from [4, 12 are a batch-size of 1024 instead of 4096, the use of global average-pooling (GAP) instead of a class token [2, 11 , fixed 2D sin-cos position embeddings [2, and the introduction of a small amount of RandAugment [3 and Mixup [21 (level 10 and probability 0.2 respectively, which is less than [12). These small changes lead to significantly better performance than that originally reported in [4.
https://arxiv.org/pdf/2205.01580.pdf ...
スピアマン相関係数
· ☕ 1 min read
todo https://webbeginner.hatenablog.com/entry/2020/06/26/120000#:~:text=2つの相関係数の違い&text=ピアソンの相関係数では%E3%80%81変数の値そのもの,順位を利用します%E3%80%82&text=正規分布に従うことを,を作ってい ...
不均衡データ
· ☕ 1 min read
reweightingはただのsoft-margin SVMと同等になるらしい ...
Attention
· ☕ 1 min read
Attentionは２つに大別される Self-Attention SourceTarget-Attention ↓ 引用 : https://www.arithmer.co.jp/post/20210413 ...
PyTorch 高速化
· ☕ 1 min read
https://qiita.com/sugulu_Ogawa_ISID/items/62f5f7adee083d96a587#31-ampautomatic-mixed-precision機能について ...
Automatic Mixed Precision
· ☕ 1 min read
float16とfloat32を混ぜて(Mixed)計算することで, GPUのMEM使用率を抑えることができる 計算スピードも幾分速くなるらしい 略してamp https://qiita.com/Sosuke115/items/40265e6aaf2e414e2fea https://tawara.hatenablog.com/entry/2021/05/31/220936 ...
Huber loss
· ☕ 1 min read
外れ値に強く, MSEよりもロバスト性が高い ...
Distribution Alignment: A Unified Framework for Long-tail Visual Recognition
· ☕ 1 min read
Decoupling Representation and Classifier for Long-Tailed Recognition を引用 新規性は以下の２つ Adaptive Calibration Function Alignment with Generalized Re-weighting Adaptive Calibration Function 分類器の出力 $\boldsymbol{z}$を線形変換して重み付け + marginを加える Alignment with Generalized Re-weighting targetの確率に重み付け https://openaccess.thecvf.com/content/CVPR2021/papers/Zhang_Distribution_Alignment_A_Unified_Framework_for_Long-Tail_Visual_Recognition_CVPR_2021_paper.pdf ...