Original Note
SapBERT - Outline
Outline
- Self-Alignment Pretraining for Biomedical Entity Representations
- To learn:
- 1. To understand list
- 2. Abstract
- 3. Related works:
- 4. Introduction:
- To understand of introduction:
- 5. Methods:
- What's the meaning of figure2, what does the author want to imply?
- To understand of methods:
- Question of formula(2):
- 6. Experiments:
- To understand of experiments:
- I skip the "The ADAPTER Variant" part
- Actual compute steps of online hard samples mining
- 第一步:计算全成对相似度矩阵 $S$
- 第二步:构建标签掩码 (Label Masks)
- 第三步:利用 $S$ 进行在线硬采样
- 第四步:计算 MS Loss (公式 2 的具体实现)