3 Matching Annotations
  1. Feb 2019
  2. iphysresearch.github.io iphysresearch.github.io
    A Paper A Day
    1
    1. Herb 01 Feb 2019
      Using Pre-Training Can Improve Model Robustness and Uncertainty

      此 paper 回应并补充了去年何神的一篇说 pre-training 对 performance 鸟用不大的文章 (Rethinking ImageNet Pre-training)。你问是怎么回应的？瞅一眼此 paper 的题目就晓得了。。。。

      pretraining robustness uncertainty
    Visit annotations in context

    Tags

    Annotators

    URL

    iphysresearch.github.io/paper_summary/APaperADay.html
  3. Nov 2018
  4. iphysresearch.github.io iphysresearch.github.io
    A Paper A Day
    2
    1. Herb 28 Nov 2018
      How Many Samples are Needed to Learn a Convolutional Neural Network?

      Conclusion 里说："Our paper only considered CNN with linear activation.“ 啥？linear activation？我不知道还有什么理由让我继续读下去[允悲]。。。另有 reddit 上对此文的讨论：http://t.cn/ELKbsjx

      pretraining
    2. Herb 23 Nov 2018
      Rethinking ImageNet Pre-training

      ImageNet预训练的反思——ImageNet预训练可在训练早起加速收敛，但未必能提供正则化或提高最终目标任务准确性

      何神的新paper！欲最快速通晓文章的 insights，请直接阅读文章的 discussions 部分即可！

      fine-tune pretraining
    Visit annotations in context

    Tags

    Annotators

    URL

    iphysresearch.github.io/paper_summary/APaperADay.html