Scholarlink

In this chapter, we discuss four articles that focus on studying properties of the training process. The first one deals with the loss function used to train a network. The other three are focused on the evolution of weights during the training process and how the fractal dimension of weight trajectories is related to the generalization capacity of a network.

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

Training Dynamics and Loss Functions

  • 章节

  • 发表日期：

    2026年

查看全文
  • Rubén Ballester,
  • Carles Casacuberta,
  • Sergio Escalera

摘要

In this chapter, we discuss four articles that focus on studying properties of the training process. The first one deals with the loss function used to train a network. The other three are focused on the evolution of weights during the training process and how the fractal dimension of weight trajectories is related to the generalization capacity of a network.