错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Scientific Analysis of ML Systems

  • Samuel Ackerman,
  • Guy Barash,
  • Eitan Farchi,
  • Orna Raz,
  • Onn Shehory

摘要

The primary requirement of a good learning system is generalization, i.e., the ability to generate rules that perform well on data similar to the training data. This is achieved through proper experimental design and the identification of governing factors. This chapter discusses how to design and use experiments in the design and implementation of ML systems.