Scholarlink

In this final chapter, we return to our study of optimization that we began in Chapter 6. Our focus here is on more advanced optimization techniques that have recently found important applications in the context of deep learning. In particular, we study momentum-based gradient descent, including the heavy ball method, Krylov subspace methods, conjugate gradients, and Nesterov acceleration, as well stochastic gradient descent.

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

Advanced Optimization

  • 章节

  • 发表日期：

    2025年

查看全文
  • Jeff Calder,
  • Peter J. Olver

摘要

In this final chapter, we return to our study of optimization that we began in Chapter 6. Our focus here is on more advanced optimization techniques that have recently found important applications in the context of deep learning. In particular, we study momentum-based gradient descent, including the heavy ball method, Krylov subspace methods, conjugate gradients, and Nesterov acceleration, as well stochastic gradient descent.