Skip to content

Data Science & Machine Learning Newsletter # 95

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group 7 Tech­niques to Han­dle Imbal­anced Data This blog post intro­duces seven tech­niques that are com­monly applied in domains like intru­sion detec­tion or real-time bid­ding, because the datasets are often extremely imbal­anced. Bayesian Deep Learn­ing with Edward (and a trick using Dropout) Ele­gant N-gram Gen­er­a­tion in Python A quick few snip­pets of code – solv­ing how to Devamını Oku […]

Data Science & Machine Learning Newsletter # 94

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Accu­rate, Large Mini­batch SGD: Train­ing Ima­geNet in 1 Hour Deep learn­ing thrives with large neural net­works and large datasets. How­ever, larger net­works and larger datasets result in longer train­ing times that impede research and devel­op­ment progress. Dis­trib­uted syn­chro­nous SGD offers a poten­tial solu­tion to this prob­lem by divid­ing SGD mini­batches over a pool of par­al­lel work­ers. … Time Devamını Oku […]

Data Science & Machine Learning Newsletter # 93

  You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Fea­ture Engi­neer­ing: Data scientist’s Secret Sauce ! It is very tempt­ing for  data sci­ence prac­ti­tion­ers to opt for the best known  algo­rithms for a given problem.However It’s not the algo­rithm alone , which can pro­vide the best solu­tion  ; Model built on care­fully engi­neered and selected fea­tures can pro­vide far bet­ter results. pix2code: Gen­er­at­ing Code from a Graph­i­cal Devamını Oku […]

Data Science & Machine Learning Newsletter # 92

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Bayesian machine learn­ing So you know the Bayes rule. How does it relate to machine learn­ing? It can be quite dif­fi­cult to grasp how the puz­zle pieces fit together — we know it took us a while. This arti­cle is an intro­duc­tion we wish we had back then. Machine Learn­ing Tech­niques for Pre­dic­tive Main­te­nance Pre­dic­tive main­te­nance pre­dicts fail­ure, and the actions could include Devamını Oku […]

Data Science & Machine Learning Newsletter # 91

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Google releases dataset of 50M vec­tor draw­ings, open sources Sketch-RNN imple­men­ta­tion Blog Post Paper Github Repo Github Repo Google Offers Cloud-Based TPU Ser­vice for Train­ing and Deploy­ing Deep Learn­ing Mod­els We’re excited to announce that our second-generation Ten­sor Pro­cess­ing Units (TPUs) are com­ing to Google Cloud to accel­er­ate a wide range of machine learn­ing Devamını Oku […]

Data Science & Machine Learning Newsletter # 90

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Sen­ti­ment analy­sis on Twit­ter using word2vec and keras word2vec is a group of Deep Learn­ing mod­els devel­oped by Google with the aim of cap­tur­ing the con­text of words while at the same time propos­ing a very effi­cient way of pre­pro­cess­ing raw text data. This post dis­c­cusses to do sen­ti­ment analy­ses with word2vec. Fully Con­vo­lu­tional Instance-aware Seman­tic Seg­men­ta­tion FCIS Devamını Oku […]

Data Science & Machine Learning Newsletter # 89

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Machine Learn­ing over­tak­ing Big Data? Big Data surged in pop­u­lar­ity both as a field and as a pop­u­lar buzz­word around 2011.   How­ever, lately is has been los­ing its lus­ter and Gart­ner in 2015 removed Big Data from its Hype Curve, replac­ing it with Machine Learn­ing. Top 10 Machine Learn­ing Videos on YouTube, updated The top machine learn­ing videos on YouTube include lec­ture Devamını Oku […]

Data Science & Machine Learning Newsletter # 88

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group How to Use Dropout with LSTM Net­works for Time Series Fore­cast­ing Long Short-Term Mem­ory (LSTM) mod­els are a type of recur­rent neural net­work capa­ble of learn­ing sequences of obser­va­tions.  This may make them a net­work well suited to time series fore­cast­ing.  An issue with LSTMs is that they can eas­ily over­fit train­ing data, reduc­ing their pre­dic­tive skill.  Dropout is a reg­u­lar­iza­tion Devamını Oku […]

Data Science & Machine Learning Newsletter # 87

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group Image-to-image trans­la­tion in PyTorch (e.g. horse2zebra, edges2cats, and more) This is an ongo­ing PyTorch imple­men­ta­tion for both unpaired and paired image-to-image trans­la­tion. Baidu’s AI just achieved Zero Shot Learn­ing As babies humans often “just learn” things, it’s a gift, but now AI’s are catch­ing us up and start­ing to learn in the same way we do and it’ll Devamını Oku […]

Data Science & Machine Learning Newsletter # 86

You want to get updates? Please join Data Sci­ence & Machine Learn­ing Newslet­ter Linked Group 10 tools and plat­forms for data prepa­ra­tion Data prepa­ra­tion has tra­di­tion­ally been a very man­ual task and con­sumed the bulk of most data project’s time.  … For­tu­nately the lat­est gen­er­a­tion of tools, typ­i­cally pow­ered by NoSQL tech­nolo­gies take a lot of this pain away. Deep Patient: An Unsu­per­vised Rep­re­sen­ta­tion to Pre­dict the Future of Patients from the Elec­tronic Health Devamını Oku […]