更新《How to find the best model parameters in scikit-learn》O網(wǎng)頁鏈接云:O網(wǎng)頁鏈接? 【視頻:DataSchool的機(jī)器學(xué)習(xí)基本概念教學(xué)】《What is machine learning, and how does it work?》O網(wǎng)頁鏈接Data School系列教程“Introduction to machine learning with scikit-learn”(基于Scikit-Learn的機(jī)器學(xué)習(xí)教程)第一講冷离,ipn:O網(wǎng)頁鏈接云:O網(wǎng)頁鏈接
黃亮也在嘗試用CNN了: [DEPENDENCY TREE-BASED CONVOLUTION FOR SENTENCE MODELING, Ma,ACL15短文]O網(wǎng)頁鏈接PS: 第二張圖是2013年8月微博評論深度學(xué)習(xí)O黃亮-算法時代? 評論指出這篇短文涉嫌抄襲北大軟工所博士生Lili Mou等人的兩篇arXiv文章雾消。Mou在學(xué)校個人主頁上的聲明:“Two of my papers are deliberately plagiarized in an ACL short paper”O網(wǎng)頁鏈接
Massively Parallel Methods for Deep Reinforcement Learning喻括,來自Google DeepmindO網(wǎng)頁鏈接
既然Torch這么叼,那貼個CVPR'15的Torch Tutorial吧。講稿連接O網(wǎng)頁鏈接
"Neural CRF Parsing"伯克利Dan Klein的NLP組ACL15文章:Neural CRF Parsing [Durrett & Klein,ACL15] 在CRF parsing中打分anchored規(guī)則產(chǎn)生式時, 將原來的基于稀疏特征的線性勢函數(shù)替換為通過前饋神經(jīng)網(wǎng)絡(luò)計(jì)算的非線性勢函數(shù). 組合稀疏的指示特征和稠密的embedding特征能進(jìn)一步提升性能. 有代碼O網(wǎng)頁鏈接?
在線實(shí)驗(yàn)室 //【想學(xué)編程嗎跛璧?先來網(wǎng)上做做實(shí)驗(yàn)吧聘芜!】O網(wǎng)頁鏈接(分享自@新聞資訊)
【適合做數(shù)據(jù)挖掘的6個經(jīng)典數(shù)據(jù)集(及另外100個列表)】《6 Useful Databases to Dig for Data (and 100 more)》O網(wǎng)頁鏈接
【視頻+講義:從實(shí)驗(yàn)室到工廠——構(gòu)建機(jī)器學(xué)習(xí)生產(chǎn)架構(gòu)】《From the Lab to the Factory: Building a Production Machine Learning Infrastructure (Cerner's Tech Talk series)》by Josh Wills from ClouderaO網(wǎng)頁鏈接云(+Slide):O網(wǎng)頁鏈接
【深度學(xué)習(xí)-LeCun、Bengio和Hinton的聯(lián)合綜述(上)】O網(wǎng)頁鏈接一個月之前的文章——深度卷積網(wǎng)絡(luò)在處理圖像侵浸、視頻、語音和音頻方面帶來了突破氛谜,而遞歸網(wǎng)絡(luò)在處理序列數(shù)據(jù)掏觉,比如文本和語音方面表現(xiàn)出了閃亮的一面。 綜述(下)鏈接:O網(wǎng)頁鏈接
工業(yè)互聯(lián)網(wǎng):突破智慧和機(jī)器的界限--GE工業(yè)互聯(lián)網(wǎng)白皮書-O網(wǎng)頁鏈接
高斯過程代碼 ?GPstuff - Gaussian process models for Bayesian analysis 4.6?O網(wǎng)頁鏈接
IJCAI-15論文集已經(jīng)可以全部開放下載了 Proceedings of the Twenty-Fourth International Joint Conference on Artificial Intelligence, Buenos Aires, Argentina, 25–31 July 2015. Edited by Qiang Yang and Michael Wooldridge, IJCAI.orgO網(wǎng)頁鏈接
這位IIT計(jì)算機(jī)科學(xué)與工程副教授的#姓名只有一個詞#Mausam值漫。研究興趣:信息抽取澳腹、常識抽取、多文檔摘要杨何、眾包和馬爾可夫決策過程酱塔。ACL15短文(錄用率22.3%): Open IE;EMNLP13: Named Entity Recognition in Tweets危虱。他也是IJCAI-15最佳論文評審委員O網(wǎng)頁鏈接
黃亮也在嘗試用CNN了: [DEPENDENCY TREE-BASED CONVOLUTION FOR SENTENCE MODELING, Ma,ACL15短文]O網(wǎng)頁鏈接PS: 第二張圖是2013年8月微博評論深度學(xué)習(xí)O黃亮-算法時代? 評論指出這篇短文涉嫌抄襲北大軟工所博士生Lili Mou等人的兩篇arXiv文章羊娃。Mou在學(xué)校個人主頁上的聲明:“Two of my papers are deliberately plagiarized in an ACL short paper”O網(wǎng)頁鏈接
【適合做數(shù)據(jù)挖掘的6個經(jīng)典數(shù)據(jù)集(及另外100個列表)】《6 Useful Databases to Dig for Data (and 100 more)》O網(wǎng)頁鏈接
【Y Bengio寫的深度學(xué)習(xí)展望文章】《The Promise of Deep Learning》By Yoshua BengioO網(wǎng)頁鏈接? 《Bengio最新博文:深度學(xué)習(xí)展望》OBengio最新博文:深度學(xué)習(xí)展望參閱O格靈深瞳
【推薦給數(shù)據(jù)科學(xué)家的七個Python工具】《Seven Python Tools All Data Scientists Should Know How to Use》IPython/GraphLab Create/Pandas/PuLP/Matplotlib/Scikit-Learn/SparkO網(wǎng)頁鏈接
【論文:在線看漲對國際金融市場影響之量化】《Quantifying the effects of online bullishness on international financial markets》H Mao, S Counts, J Bollen (2015)O網(wǎng)頁鏈接相關(guān)報道《If Twitter Is Bullish, Maybe You Should Be Too》O網(wǎng)頁鏈接
【開源:(Java)在線Twitter情感判斷Umigon】O網(wǎng)頁鏈接Paper:O網(wǎng)頁鏈接GitHub:O網(wǎng)頁鏈接情感判斷規(guī)則(表):O網(wǎng)頁鏈接
【講義:Aaron Courville的RNN教程】《Recurrent Neural Networks》by Aaron CourvilleO網(wǎng)頁鏈接選自"IFT6266 – H2015 Representation Learning"茸苇,請參閱O愛可可-愛生活
【又一篇ICML 2015回顧】《ICML 2015 Review》by Paul MineiroO網(wǎng)頁鏈接
【Twitter2009年全量數(shù)據(jù)集】《Twitter Social Graph 2009》"41.7 million user profiles, 1.47 billion social relations, 4,262 trending topics, and 106 million tweets"O網(wǎng)頁鏈接
【Scaling Up Machine Learning: Parallel and Distributed Approaches】Scaling Up Machine Learning: Parallel and Distributed ApproachesO網(wǎng)頁鏈接【Introduction:O網(wǎng)頁鏈接】 【Tree ensembles:O網(wǎng)頁鏈接】 【Graphical models:O網(wǎng)頁鏈接】 【Summary】
【殺入Kaggle機(jī)器人檢測競賽前十的經(jīng)驗(yàn)分享】《Getting into Top 10 in Kaggle Facebook Recruiting Competition》O網(wǎng)頁鏈接
【視頻:DataSchool的機(jī)器學(xué)習(xí)基本概念教學(xué)】《What is machine learning, and how does it work?》O網(wǎng)頁鏈接Data School系列教程“Introduction to machine learning with scikit-learn”(基于Scikit-Learn的機(jī)器學(xué)習(xí)教程)第一講姐霍,ipn:O網(wǎng)頁鏈接云:O網(wǎng)頁鏈接? 更新《How to find the best model parameters in scikit-learn》O網(wǎng)頁鏈接云:O網(wǎng)頁鏈接
【視頻:不依賴(大量)標(biāo)注的文本分類——詞向量應(yīng)用】《Text By the Bay 2015: Mike Tamir, Classifying Text without (many) Labels》O網(wǎng)頁鏈接云:O網(wǎng)頁鏈接
【視頻+講義:(來自DL4J創(chuàng)始人Adam Gibson)深度學(xué)習(xí)及其機(jī)器視覺應(yīng)用】《Practical Deep Learning and Its Applications: Computer Vision》by Adam GibsonO網(wǎng)頁鏈接云(+Slide):O網(wǎng)頁鏈接
【視頻+講義:(ILSVRC2014)Google面向機(jī)器視覺的深度學(xué)習(xí)】《Deep Networks for Computer Vision at Google – ILSVRC2014》O網(wǎng)頁鏈接云:O網(wǎng)頁鏈接Slide:O網(wǎng)頁鏈接
PU(positive-unlabeled) Learning for Matrix Completion [Hsieh et al,ICML'15]O網(wǎng)頁鏈接PU學(xué)習(xí)的應(yīng)用場景包括推薦系統(tǒng)和社會網(wǎng)絡(luò)中僅有【正類】可觀察實(shí)例(如點(diǎn)贊和好友關(guān)系),剩下的是未標(biāo)注的實(shí)例,沒有負(fù)類實(shí)例. 實(shí)驗(yàn)評估有鏈接預(yù)測. PS:arXiv'14的第2作者被替換為ICML'15上的第2作者? PU-learning, B.Liu: 1) ICML02:Partially Supervised Classification; ICML03:Weighted Logistic Regression; IJCAI03:Rocchio & SVM; ICDM03:biased SVM; ECML05:Different Distributions 2) EMNLP10:Negative; ACL10:DistSim vs PUO網(wǎng)頁鏈接
浙江大學(xué)CAD重點(diǎn)實(shí)驗(yàn)室“計(jì)算機(jī)圖形學(xué)與大規(guī)模數(shù)據(jù)分析”的課件:O網(wǎng)頁鏈接。涉及計(jì)算機(jī)圖像圖形學(xué)扣讼、大數(shù)據(jù)弥雹、可視化垃帅、機(jī)器學(xué)習(xí)等,例如:真實(shí)感圖形繪制缅糟;跨媒體理解中的結(jié)構(gòu)性學(xué)習(xí)挺智;可視化技術(shù)成就淘寶數(shù)據(jù)之美-賈超(淘寶);大數(shù)據(jù)的智能處理窗宦;基于物理的計(jì)算機(jī)動畫赦颇;流行學(xué)習(xí)等等
IJCAI-15論文集已經(jīng)可以全部開放下載了 Proceedings of the Twenty-Fourth International Joint Conference on Artificial Intelligence, Buenos Aires, Argentina, 25–31 July 2015. Edited by Qiang Yang and Michael Wooldridge, IJCAI.orgO網(wǎng)頁鏈接
得益于現(xiàn)在更為強(qiáng)大的計(jì)算機(jī)、可用的海量豐富數(shù)據(jù)集以及先進(jìn)的算法赴涵,我們終于可以跨越一個長期以來阻礙計(jì)算機(jī)科學(xué)發(fā)展的閾值媒怯。機(jī)器學(xué)習(xí)正在從一個高度人工化的階段向另一個更為自動化的階段進(jìn)行快速轉(zhuǎn)變。OBengio最新博文:深度學(xué)習(xí)展望
高斯過程代碼髓窜, GPstuff - Gaussian process models for Bayesian analysis 4.6O網(wǎng)頁鏈接
A Step by Step Backpropagation Example#反向傳播BP算法教程#…O網(wǎng)頁鏈接
Gaussian Process Summer School, 2015O網(wǎng)頁鏈接
From Autoencoders to Autoregressive Models (Masked Autoencoders ICML Paper)O網(wǎng)頁鏈接