【論文メモ】Rethinking the Value of Labels for Improving Class-Imbalanced Learning
· ☕ 1 min read
不均衡データには正と負の両方の側面がある 正の側面 性能に寄与する → 負の側面 サンプル数が多いクラスに引っ張られて決定境界が歪む → https://arxiv.org/abs/2006.07529 ...
todo
【論文メモ】No Parameters Left Behind: Sensitivity Guided Adaptive Learning Rate for Training Large Transformer Models
· ☕ 1 min read
https://arxiv.org/pdf/2202.02664.pdf ...
【論文メモ】Should You Mask 15% in Masked Language Modeling?
· ☕ 1 min read
https://arxiv.org/pdf/2202.08005.pdf ...
【論文メモ】Self-Supervised Learning for Semi-Supervised Time Series Classification
· ☕ 1 min read
https://www.ismll.uni-hildesheim.de/pub/pdfs/pakdd_shayan.pdf ...