04 2018 档案
摘要：in policy gradient, "a" is replaced by "u" usually. use this new form to estimate how good the update is. If all three path show positive reward, shou
阅读全文
摘要：https://www.youtube.com/watch?v=fevMOp5TDQs http://www.denizyuret.com/2015/03/alec-radfords-animations-for.html artari is not a MDP, but MDP method wo
阅读全文
摘要：...
阅读全文
摘要：https://www.youtube.com/watch?v=qaMdN6LS9rA https://drive.google.com/file/d/0BxXI_RttTZAhVXBlMUVkQ1BVVDQ/view match: a4 b1 c2 d3 a The middle one is c
阅读全文
摘要：T in Hinton's paper, was set to 10 initially, and went done gradually to zero.
阅读全文
摘要：\
阅读全文
摘要：symmetric version: called Hopfield Net if y <-- minus y, energy won't change! the second last line's first "-" is not the minus sign https://zhuanlan.
阅读全文
摘要：https://en.wikipedia.org/wiki/Liquid_state_machine https://www.youtube.com/watch?v=r4wxN2B5PuU https://stackoverflow.com/questions/28326776/liquid-sta
阅读全文
摘要：http://deeplearning.cs.cmu.edu/recitations/rec8.attention.pdf https://github.com/cmudeeplearning11785/deep-learning-tutorials/blob/master/recitation-8
阅读全文
摘要：https://www.youtube.com/watch?v=oiNFCbD_4Tk <eos> end of section a common trick image to seq problem mistake might happen. look at the fig(1,0) could
阅读全文
摘要：denoise autoencoder AE generates things that look like trained data 太难了，学不下去了！准备zs 学不下去了！
阅读全文
摘要：NN is pretty bad at learning this pattern. green dots are referred to the first layer. blue -> second layer red -> third layer https://github.com/cmud
阅读全文
摘要：https://distill.pub/2017/ctc/ https://en.wikipedia.org/wiki/Beam_search
阅读全文
摘要：...
阅读全文
摘要：Xent = cross entropy
阅读全文
摘要：https://github.com/cmudeeplearning11785/deep-learning-tutorials/blob/master/recitation-5/recitation5.ipynb http://colah.github.io/posts/2015-08-Unders
阅读全文
摘要：Div(Y(1),...,Y(t),Ytarget(1),...,Ytarget(t)) ≠ ∑Div(Y(i),Ytarget(i)) Recurrence is not necessary. http://www.hexahedria.com/2015/08/03/composing-music
阅读全文
摘要：What's the meaning of the forth equation???? What is Hk?? http://colah.github.io/posts/2015-08-Understanding-LSTMs/
阅读全文
摘要：http://deeplearning.cs.cmu.edu/slides/lec11.recurrent.pdf I think the subscripts in this lecture is quite confusing, and even incorrect sometimes. Jac
阅读全文
摘要：In linear algebra class, the cross product of matrix is not mainly discussed, and actually, as we all known, cross product will introduce a new dimens
阅读全文
摘要：...
阅读全文
摘要：https://zhuanlan.zhihu.com/p/22810533 L2 -> Regression Problems KL -> Classification Problems http://deeplearning.cs.cmu.edu/slides/lec8.stochastic_gr
阅读全文
摘要：http://deeplearning.cs.cmu.edu/slides/lec7.cascor.pdf this is the guest lecture with no video. So let's ignore it. Recitation 4: CNN ' https://github.
阅读全文
摘要：Question: Does that mean that former choosed data is more important than subsequent data?
阅读全文
摘要：
阅读全文
摘要：https://github.com/cmudeeplearning11785/deep-learning-tutorials/blob/master/recitation-3/recitation3.ipynb https://blog.csdn.net/shuzfan/article/detai
阅读全文
摘要：
阅读全文
摘要：
阅读全文
摘要：Recitation 2 numpy operations array index x = np.arange(10) 2 x:[ 0 1 4 9 16 25 36 49 64 81] print(x[:: 1]) all reversed print(x[8:1: 1]) reversed sli
阅读全文
摘要：4/16/2018 前段时间自己动手实践了一下不带DNN的强化学习算法，今天正式开始学DNN 算了，这个课程太基础了，也就是把Machine Learning里NN的部分稍微多讲了一点。
阅读全文
摘要：Doctor of Philosophy (PhD) The PhD requires a minimum academic residency of two years beyond the bachelor’s degree. Programs are individually develope
阅读全文
摘要：Laboratory on Reinforcement Learning and Decision Making PSY 2350R Samuel Gershman 2017 Fall Full Term 8/30/2017 to 12/1/2017 S M T W Th F S 10:00am -
阅读全文
摘要：对网上各处寻找reinforcement learning的实践环节感到无力，尤其是对open ai gym的介绍上。cnblogs、github、csdn、zhuanlan.zhihu、公众号、google、youtube等等，都非常的零碎。公开课上主要是理论，对代码实现讲的很少。对于像我这样的新
阅读全文
摘要：因为需要用gym，所以不得不用linux 考虑性能和易用性，因此不想用虚拟机和wsl，安装ubuntu各种坑，踩了3天，总算完成各种部署了！ 任务完成：python、anaconda、tensorflow-gpu、vscode 1，ubuntu17.10下载，rufus写入U盘，关闭fast boo
阅读全文
浙公网安备 33010602011771号