01 2019 档案

2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 25} [Spectral Methods]
摘要：draw a topic h and then draw independent xfrom multinomial distribution A paper can have several topics, and each topics has different proportions O i 阅读全文
posted @ 2019-01-17 12:59 ecoflex 阅读(170) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 23} [Applications in Computer Vision (cont’d) + Gaussian Process] (unfinished)
摘要：(unfinished) 阅读全文
posted @ 2019-01-14 09:44 ecoflex 阅读(116) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 22} [Applications in Computer Vision (cont’d) + Gaussian Process]
摘要：take log to Normal distribution, it's L2 loss, which might be the cause of the blurry results of VAE interpreting the meaning of latent vars is diffic 阅读全文
posted @ 2019-01-14 08:44 ecoflex 阅读(252) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 21} [A Hybrid: Deep Learning and Graphical Models]
摘要：conent in green box difficult to solve https://www.youtube.com/watch?v=btr1poCYIzw&t=60s 阅读全文
posted @ 2019-01-14 07:04 ecoflex 阅读(193) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 15} [Mean field Approximation]
摘要：p_bar(x) = p(x,e) p(x) is p(x|e) p_bar is hard to estimate, but p(x) is easy and p(x) \propotion p_bar q is easier for sampling, while we cant sample 阅读全文
posted @ 2019-01-12 13:13 ecoflex 阅读(311) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 14} [Loopy Belief Propagation] (unfinished)
摘要： 阅读全文
posted @ 2019-01-10 12:49 ecoflex 阅读(318) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 11} [CRF (Cont'd) + Intro to Topic Models]
摘要： 阅读全文
posted @ 2019-01-07 09:29 ecoflex 阅读(218) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 10} [HHM and CRF]
摘要：between tags and words, there's table 1. between tags, there's table 2. combine the two tables, p(...) to get the results. MRF: factors of the tables 阅读全文
posted @ 2019-01-07 08:22 ecoflex 阅读(239) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 6} [Factor graph, message passing, and Junction Tree] （unfinished）
摘要： 阅读全文
posted @ 2019-01-07 06:11 ecoflex 阅读(235) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 5} [Algorithms for Exact Inference]
摘要：Not in the frontier of research, but the results are used commonly now. X_{k+1} - X_n are known, to calculate the joint probability, we have to do inf 阅读全文
posted @ 2019-01-07 04:30 ecoflex 阅读(223) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 3} [Undirected Graphic Model]
摘要：DGM ignores sometimes the latent variable know parts of the proteins' states, and the corr relationships, how would other proteins's state probably be 阅读全文
posted @ 2019-01-06 08:34 ecoflex 阅读(251) 评论(0) 推荐(0)
2018 10-708 (CMU) Probabilistic Graphical Models {Lecture 2} [Directed GMs: Bayesian Networks]
摘要：https://kayhan.dbmi.pitt.edu/sites/default/files/lecture2.pdf P(sequence) given both dices are fair x2 → x4 ↗ ↘ x1 x5 ↘ ↗ x3 P(x5) = P(X1) P(X2|X1) P( 阅读全文
posted @ 2019-01-05 15:11 ecoflex 阅读(229) 评论(0) 推荐(0)
Probabilistic Graphical Models 10-708, Spring 2017
摘要：https://www.cs.cmu.edu/~epxing/Class/10708-17/slides/lecture1-Introduction.pdf Computational and CS orientated => DK and NF's book Statistical and eas 阅读全文
posted @ 2019-01-04 09:40 ecoflex 阅读(222) 评论(0) 推荐(0)
Modern C++ Course [Lecture 9] {Templates, Iterators, Exceptions, Program input parameters, OpenCV}
摘要：GP is a kind of programming that will generate code for you at compile time. The core is to seperate algorithms from the data structure, and we do som 阅读全文
posted @ 2019-01-03 14:24 ecoflex 阅读(165) 评论(0) 推荐(0)
如何在macOS mojave上安装opencv4给vscode使用
摘要：first, install opcv4, the newest one. but brew install opencv will only get a version 3, Another way is to use source code and compile it locally, but 阅读全文
posted @ 2019-01-03 13:46 ecoflex 阅读(950) 评论(0) 推荐(0)
Modern C++ Course [Lecture 8] {Smart/Unique/Shared pointers, Associative containers, Type casting, Enumeration, Binary files}
摘要：only use them with heap memory. never ever apply them to stack try to avoid using "new and delete" if you can use shared or unique ptr delete the a_pt 阅读全文
posted @ 2019-01-03 02:47 ecoflex 阅读(166) 评论(0) 推荐(0)
Modern C++ Course [Lecture 7] {Pointers, const with pointers, Stack and Heap, Memory leaks, Dangling pointers}
摘要：https://en.cppreference.com/w/cpp/language/range-for every object has a pointer to itsleft. stack operations are very quick it's computationally expen 阅读全文
posted @ 2019-01-01 11:40 ecoflex 阅读(135) 评论(0) 推荐(0)