avatar
Articles
200
Tags
38
Categories
0

Home
Archives
Tags
Categories
List
  • Music
  • Movie
Link
About
ALTNT's Hexo Blog
Home
Archives
Tags
Categories
List
  • Music
  • Movie
Link
About

Does Negative Sampling Matter? A Review with Insights into its Theory and Applications

Created2025-09-01|Updated2025-09-05
|Post Views:
Author: ALTNT
Link: http://blog.705553939.xyz/2025/09/01/sequence-processing/2024-NegativeSampling/
Copyright Notice: All articles in this blog are licensed under CC BY-NC-SA 4.0 unless stating additionally.
Sequence Processing
Previous
周报2025年9月7日
Next
周报2025年8月31日
Related Articles
2024-11-25
CSDI:Conditional Score-based Diffusion Models for Probabilistic Time Series Imputation
2024-12-08
Representation Learning with Contrastive Predictive Coding
2024-11-24
CSDI代码解释
2025-05-01
Differential Transformer
2025-04-10
Improving Time Series Encoding with Noise-Aware Self-Supervised Learning and an Efficient Encoder
avatar
ALTNT
Articles
200
Tags
38
Categories
0
Follow Me
Announcement
This is my Blog
Contents
  1. 1. Does Negative Sampling Matter? A Review with Insights into its Theory and Applications
    1. 1.1. 摘要
    2. 1.2. 1 引言
      1. 1.2.1. 主要贡献
      2. 1.2.2. 综述结构
      3. 1.2.3. 变量定义
      4. 1.2.4. 术语说明(便于领域理解)
  2. 2. 2 负采样
    1. 2.1. 2.1 形式化定义与框架
    2. 2.2. 2.2 负采样的发展历程
      1. 2.2.0.1. (随机负采样)
    3. 2.2.1. (基于流行度的负采样(PNS))
    4. 2.2.2. 难负采样(Hard NS)[19,33,34,35]策略
    5. 2.2.3. 基于GAN的负采样(GAN-based NS)[9,43,44,45,46,47]技术
    6. 2.2.4. 批内负采样(In-batch NS)策略
  3. 2.3. 2.3 负采样的重要性
    1. 2.3.1. 1)计算效率
    2. 2.3.2. 2)处理类别不平衡问题
    3. 2.3.3. 3)提升模型性能
  4. 2.4. 3 算法
    1. 2.4.1. 3.1 负样本候选集构建
      1. 2.4.1.1. 3.1.1 全局选择(Global Selection)
      2. 2.4.1.2. 3.1.2 局部选择(Local Selection)
      3. 2.4.1.3. 3.1.3 批内选择(Mini-batch Selection)
      4. 2.4.1.4. 3.1.4 跳数选择(Hop Selection)
      5. 2.4.1.5. 3.1.5 基于内存的选择(Memory-based Selection)
    2. 2.4.2. 3.2 负采样算法
      1. 2.4.2.1. 3.2.1 静态负采样(Static Negative Sampling)
        1. 2.4.2.1.1. (1)随机负采样(Random Negative Sampling, RNS)
        2. 2.4.2.1.2. (2)基于流行度的负采样(Popularity-based Negative Sampling, PNS)
        3. 2.4.2.1.3. 优缺点
      2. 2.4.2.2. 3.2.2 难负采样(Hard Negative Sampling)
        1. 2.4.2.2.1. (1)动态负采样(Dynamic Negative Sampling, DNS)
        2. 2.4.2.2.2. (2)基于混合的动态负采样(Mixture-based DNS)
        3. 2.4.2.2.3. 优缺点
      3. 2.4.2.3. 3.2.3 基于生成对抗网络的负采样(GAN-based Negative Sampling)
        1. 2.4.2.3.1. 基于生成对抗网络的负样本挖掘(GAN-based Negative Mining)
        2. 2.4.2.3.2. 基于生成对抗网络的负样本生成(GAN-based Negative Generation)
        3. 2.4.2.3.3. 优缺点
      4. 2.4.2.4. 3.2.4 基于辅助信息的负采样(Auxiliary-based Negative Sampling)
        1. 2.4.2.4.1. 基于额外数据的负采样(Extra-data-based NS)
        2. 2.4.2.4.2. 基于图的负采样(Graph-based NS)
        3. 2.4.2.4.3. 基于缓存的负采样(Cache-based NS)
        4. 2.4.2.4.4. 优缺点
      5. 2.4.2.5. 3.2.5 批内负采样(In-batch Negative Sampling)
        1. 2.4.2.5.1. 基础型批内负采样(Basic In-batch NS)
        2. 2.4.2.5.2. 去偏型批内负采样(Debiased In-batch NS)
        3. 2.4.2.5.3. 硬负型批内负采样(Hard In-batch NS)
        4. 2.4.2.5.4. 优缺点
  5. 2.5. 4 应用领域
    1. 2.5.1. 4.1 负采样在推荐系统中的应用
    2. 2.5.2. 4.2 负采样在图表示学习中的应用
    3. 2.5.3. 4.3 负采样在知识图谱嵌入中的应用
    4. 2.5.4. 4.4 负采样在自然语言处理中的应用
    5. 2.5.5. 4.5 负采样在计算机视觉中的应用
  6. 2.6. 5 讨论与未来方向
    1. 2.6.1. 5.1 无采样策略(Non-Sampling)
    2. 2.6.2. 5.2 摆脱负采样(Getting Rid of Negative Sampling)
    3. 2.6.3. 5.3 负样本数量(The Quantity of Negative Samples)
    4. 2.6.4. 5.4 负样本质量(The Quality of Negative Samples)
    5. 2.6.5. 5.5 假负样本问题(False Negative Issue)
  7. 2.7. 6 结论
  8. 2.8. 一、研究背景与核心问题
  9. 2.9. 二、负采样基础理论
    1. 2.9.1. 1. 典型应用案例:word2vec的Skip-Gram模型
    2. 2.9.2. 2. 通用框架与形式化定义
    3. 2.9.3. 3. 发展历程(五条演进路径)
  10. 2.10. 三、负采样关键技术
    1. 2.10.1. 1. 负样本候选池构建(“从哪里采样”)
    2. 2.10.2. 2. 负采样算法分类(“如何采样”)
  11. 2.11. 四、负采样的应用领域
    1. 2.11.1. 1. 推荐系统(RS)
    2. 2.11.2. 2. 图表示学习(GRL)
    3. 2.11.3. 3. 知识图谱嵌入(KGE)
    4. 2.11.4. 4. 自然语言处理(NLP)
    5. 2.11.5. 5. 计算机视觉(CV)
  12. 2.12. 五、开放问题与未来方向
  13. 2.13. 六、核心结论
Recent Post
周报2025年12月7日2025-12-08
周报2025年11月30日2025-11-30
LP-FT2025-11-23
周报2025年11月22日2025-11-22
In-season crop progress in unsurveyed regions using networks trained on synthetic data2025-11-16
©2020 - 2025 By ALTNT
Framework Hexo|Theme Butterfly