随笔分类 -  科研进程

Incremental learning
摘要：1. what is incremental learning? incremental learning is a method of machine learning in which input data is continuously used to extend the existing 阅读全文
posted @ 2021-03-08 10:16 keeps_you_warm 阅读(120) 评论(0) 推荐(0) 编辑
how to write a paper - logic part (2)
摘要：1. know your data. 知道里面有哪些pattern，正常的pattern是什么，不正常的pattern是什么，点异常的情况是什么，模式异常的情况是什么。 异常： 已知 可以通过预处理处理掉的 不能通过预处理处理掉的，如何在loss部分进行处理。 未知 motif: 小的、细微的频繁出 阅读全文
posted @ 2020-08-04 17:46 keeps_you_warm 阅读(102) 评论(0) 推荐(0) 编辑
how to write a paper - logical part(1)
摘要：像AAAI，IJCAI这种水会，投论文更多的像是掷筛子，随机性很大，你投不中不能说明你不好，你投中了也不能说明你强。 从审稿人角度考虑，如果这篇文章他看得懂，那他很可能会拒掉；如果他看不懂，那么他觉得这篇文章可能有价值，得琢磨一下。 纯方法类的文章研究和提升起来是很难的，很少有人做这个，所以作为普通 阅读全文
posted @ 2020-08-01 21:56 keeps_you_warm 阅读(149) 评论(0) 推荐(1) 编辑
attention mechanism思考
摘要：1. 了解有几种attention mechanism Seq2Seq, from a sequence input to a sequence output. Align & Translate, A potential problem of the vanilla Seq2Seq archite 阅读全文
posted @ 2020-07-29 03:34 keeps_you_warm 阅读(163) 评论(0) 推荐(0) 编辑
写作技巧
摘要：2思考问题，写introduction的时候，是用领域问题描述，还是通用方法，如果数据特殊就说领域问题，但你需要指出这类领域数据的特点和scientific challenge.，比如能源的异常检测和其他的有啥区别，为什么传统方法不行。 通用方法创新要难一些。 introduction两方面，mot 阅读全文
posted @ 2020-07-22 00:09 keeps_you_warm 阅读(129) 评论(0) 推荐(1) 编辑
PP: Meta-learning framework with applications to zero-shot time-series forecasting
摘要：From: Yoshua Bengio Problem: time series forecasting. Supplementary knowledge: 1. what is meta-learning: https://www.zhihu.com/question/264595128 2. w 阅读全文
posted @ 2020-02-14 22:09 keeps_you_warm 阅读(396) 评论(0) 推荐(0) 编辑
Conference deadlines
摘要：1. NLP/IR/DM/ML Conference Deadlines(Updating) Two Principles of Deadlines:1. All deadlines converge on the same day—Deadline Day.2. Every day is Dead 阅读全文
posted @ 2020-02-14 18:25 keeps_you_warm 阅读(1294) 评论(0) 推荐(1) 编辑
PP: Articial Intelligence—TheRevolution Hasn’t HappenedYet
摘要：From Michael I. Jordan As with many phrases that cross over from technical academic fields into general circulation, there is significant misunderstan 阅读全文
posted @ 2020-02-12 17:42 keeps_you_warm 阅读(227) 评论(0) 推荐(0) 编辑
PP: Taking the human out of the loop: A review of bayesian optimization
摘要：Problem: Design problem parameters consist of the search space of your model. Scientists design experiments to gain insights into physical and social 阅读全文
posted @ 2020-02-12 02:26 keeps_you_warm 阅读(340) 评论(0) 推荐(0) 编辑
Normalizing flows
摘要：probability VS likelihood: https://zhuanlan.zhihu.com/p/25768606 http://sdsy888.me/%E9%9A%8F%E7%AC%94-Writing/2018/%E4%BC%BC%E7%84%B6%EF%BC%88likeliho 阅读全文
posted @ 2020-02-12 01:26 keeps_you_warm 阅读(244) 评论(0) 推荐(0) 编辑
PP: Neural ordinary differential equations
摘要：Instead of specifying a discrete sequence of hidden layers, we parameterize the derivative of the hidden state using a neural network. Before: a discr 阅读全文
posted @ 2020-02-12 01:21 keeps_you_warm 阅读(226) 评论(0) 推荐(0) 编辑
AI: Uninformed search
摘要：What is a search problem: A solution to a search problem is a sequence of actions (a path) from s0 to a goal state. It is optimal if it has minimum su 阅读全文
posted @ 2020-02-11 15:52 keeps_you_warm 阅读(248) 评论(0) 推荐(0) 编辑
BK: Data mining, Chapter 2 - getting to know your data
摘要：Why: real-world data are typically noisy, enormous in volume, and may originate from a hodgepodge of heterogeneous sources. mean; median; mode(most co 阅读全文
posted @ 2020-02-11 02:23 keeps_you_warm 阅读(142) 评论(0) 推荐(0) 编辑

点击右上角即可分享
微信分享提示