Infinite Latent Feature Models and the Indian Buffet Process - paperbox @ ウィキ

Infinite Latent Feature Models and the Indian Buffet Process

• abstract
• Introduction
• A distribution on infinite binary matrices
  • A finite feature model
  • Ewuivalence classes
  • Taking the infinite limit
  • The Indian buffet process
  • Conditional distributions
• A liner-Gaussian binary latent feature model
• Conclusion

CRPはcustomerがalignできるtable(=素性)は１つ
でもIBPではdish(=素性)を取り放題ですよーみたいな確率過程
でもまだよく分からないから、読み込もう

やっぱ違うのにするかも