**The academic tip: What is Deep Learning?**

**TensorFlow Tutorial – Simple Linear Model**

Advertisements

**23**
*Monday*
Oct 2017

Posted Distilled News

in**The academic tip: What is Deep Learning?**

The commonly called deep learning or hierarchical learning is now a popular trend in machine learning. Recently during the Swiss Analytics Meeting Prof. Dr. Sven F. Crone presented how we can use deep learning in the industry in a forecasting perspective (beer forecasting for manufacturing, lettuce forecasting in retail outlets, container forecasts). Deep learning has a variety of applications as for example image and handwritten character recognition. It analyses a picture and will be able to conclude if it is a dog, a human or something else. After a learning process, deep learning first understands your handwriting and then can read and interpret a draft paper you have quickly written. But briefly what is exactly deep learning? In the artificial intelligence process, deep learning plays an important role. It is considered as a method of machine learning and roughly speaking means neural networks. More precisely artificial neural networks are intended to simulate the behaviour of biological systems composed of multiple layers of nodes (or computational units), usually interconnected in feed-forward way. Each node in one layer has directed connections to the nodes of the subsequent layer. Feed-forward neural networks can be considered as a type of non-linear predictive models that takes inputs (very often huge amount of both labelled and unlabelled data), transforms and weights these through plenty of hidden layers to produce a set of outputs (predictions). The use of a sequence of layers, organised in deep or hierarchical levels, explains the term of « deep learning ». Each layer receives as input the information contained in the previous layer, transforms it to the following layer and of course complete and improve it.

**TensorFlow Tutorial – Simple Linear Model**

In this excellent tutorial video presentation below, Magnus Erik Hvass Pedersen demonstrates the basic workflow of using TensorFlow with a simple linear model. After loading the so-called MNIST data-set with images of hand-written digits, he defines and optimizes a simple mathematical model in TensorFlow. The results are then plotted and discussed. You can find the Jupyter notebook for the demo HERE. You should be familiar with basic linear algebra, Python and the Jupyter Notebook editor. It also helps if you have a basic understanding of Machine Learning and classification.

Advertisements

**23**
*Monday*
Oct 2017

Posted R Packages

in* Multivariate Symmetric Uncertainty and Other Measurements* (

Estimators for multivariate symmetrical uncertainty based on the work of Gustavo Sosa et al. (2016) <arXiv:1709.08730>, total correlation, information gain and symmetrical uncertainty of categorical variables.

Implementation of higher order generalized singular value decomposition (HO GSVD). Based on Ponnapalli, Saunders, etal (2011) <doi:10.1371/journal.pone.0028072>.

Power of non-parametric Mann-Kendall test is highly influenced by serially correlated data. To address this issue, original time-series is modified by removing any trend component existing in the data and calculating effective sample size. Hamed, K. H., & Ramachandra Rao, A. (1998). A modified Mann-Kendall trend test for auto correlated data. Journal of Hydrology, 204(1-4), 182-196. <doi:10.1016/S0022-1694(97)00125-X>. Yue, S., & Wang, C. Y. (2004). The Mann-Kendall test modified by effective sample size to detect trend in serially correlated hydrological series. Water Resources Management, 18(3), 201-218. <doi:10.1023/B:WARM.0000043140.61082.60>.

Accomplishes mark-recapture analysis with covariates. Models available include the Cormack-Jolly-Seber open population (Cormack (1972) <doi:10.2307/2556151>; Jolly (1965) <doi:10.2307/2333826>; Seber (1965) <doi:10.2307/2333827>) and Huggin’s (1989) <doi:10.2307/2336377> closed population. Link functions include logit, sine, and hazard. Model selection, model averaging, plot, and simulation routines included. Open population size by the Horvitz-Thompson (1959) <doi:10.2307/2280784> estimator.

Different examples and methods for testing (including different proposals described in Ameijeiras-Alonso et al., 2016 <arXiv:1609.05188>) and exploring (including the mode tree, mode forest and SiZer) the number of modes using nonparametric techniques.

When security risks in web services are discovered by independent security researchers who understand the severity of the risk, they often lack the channels to properly disclose them. As a result, security issues may be left unreported. The ‘security.txt’ ‘Web Security Policies’ specification defines an ‘IETF’ draft standard <https://…/draft-foudil-securitytxt-00> to help organizations define the process for security researchers to securely disclose security vulnerabilities. Tools are provided to help identify and parse ‘security.txt’ files to enable analysis of the usage and adoption of these policies.

**23**
*Monday*
Oct 2017

Posted Books

in
**23**
*Monday*
Oct 2017

Posted Magister Dixit

in“Improving Visual Data Discovery:

1. Always have new data sources.

2. Always have new techniques.

3. Always have new tools and platforms.

Visual data discovery is not once and done. It is an iterative process that requires communication and exploration.” Analise Polsky ( 2014 )

**22**
*Sunday*
Oct 2017

Posted Documents

inAlgorithms which compute properties over graphs have always been of interest in computer science, with some of the fundamental algorithms, such as Dijkstra’s algorithm, dating back to the 50s. Since the 70s there as been interest in computing over graphs which are constantly changing, in a way which is more efficient than simple recomputing after each time the graph changes. In this paper we provide a survey of both the foundational, and the state of the art, algorithms which solve either shortest path or transitive closure problems in either fully or partially dynamic graphs. We balance this with the known conditional lowerbounds. Dynamic Shortest Path and Transitive Closure Algorithms: A Survey

**22**
*Sunday*
Oct 2017

Posted What is ...

in**Fader Network**

This paper introduces a new encoder-decoder architecture that is trained to reconstruct images by disentangling the salient information of the image and the values of attributes directly in the latent space. As a result, after training, our model can generate different realistic versions of an input image by varying the attribute values. By using continuous attribute values, we can choose how much a specific attribute is perceivable in the generated image. This property could allow for applications where users can modify an image using sliding knobs, like faders on a mixing console, to change the facial expression of a portrait, or to update the color of some objects. Compared to the state-of-the-art which mostly relies on training adversarial networks in pixel space by altering attribute values at train time, our approach results in much simpler training schemes and nicely scales to multiple attributes. We present evidence that our model can significantly change the perceived value of the attributes while preserving the naturalness of images. … **Deep Ritz Method**

We propose a deep learning based method, the Deep Ritz Method, for numerically solving variational problems, particularly the ones that arise from partial differential equations. The Deep Ritz method is naturally nonlinear, naturally adaptive and has the potential to work in rather high dimensions. The framework is quite simple and fits well with the stochastic gradient descent method used in deep learning. We illustrate the method on several problems including some eigenvalue problems. … **rApache**

rApache is a project supporting web application development using the R statistical language and environment and the Apache web server. The current software distribution runs on UNIX/Linux and Mac OS X operating systems. Apache servers with threaded Multi-Processing Modules are now supported, but the the Apache Prefork Multi-Processing Module is still recommended (refer to the Multi-Processing Modules chapter from Apache for more about this). The rApache software distribution provides the Apache module named mod_R that embeds the R interpreter inside the web server. It also comes bundled with libapreq, an Apache module for manipulating client request data. Together, they provide the glue to transform R into a server-side scripting environment. Another important project that’s not bundled with rApache, but plays an important role in server-side scripting, is the R package brew (also available on CRAN). It implements a templating framework for report generation, and it’s perfect for generating HTML on the fly. it’s syntax is similar to PHP, Ruby’s erb module, Java Server Pages, and Python’s psp module. brew can be used stand-alone as well, so it’s not part of the distribution.

http://…/rscript-as-service-api …

**22**
*Sunday*
Oct 2017

Posted R Packages

in* Nested Loop Cross Validation* (

Nested loop cross validation for classification purposes for misclassification error rate estimation. The package supports several methodologies for feature selection: random forest, Student t-test, limma, and provides an interface to the following classification methods in the ‘MLInterfaces’ package: linear, quadratic discriminant analyses, random forest, bagging, prediction analysis for microarray, generalized linear model, support vector machine (svm and ksvm). Visualizations to assess the quality of the classifier are included: plot of the ranks of the features, scores plot for a specific classification algorithm and number of features, misclassification rate for the different number of features and classification algorithms tested and ROC plot. For further details about the methodology, please check: Markus Ruschhaupt, Wolfgang Huber, Annemarie Poustka, and Ulrich Mansmann (2004) <doi:10.2202/1544-6115.1078>.

Obtain estimated marginal means (EMMs) for many linear, generalized linear, and mixed models. Compute contrasts or linear functions of EMMs, trends, and comparisons of slopes. Plots and compact letter displays. Least-squares means are discussed, and the term ‘estimated marginal means’ is suggested, in Searle, Speed, and Milliken (1980) Population marginal means in the linear model: An alternative to least squares means, The American Statistician 34(4), 216-221 <doi:10.1080/00031305.1980.10483031>.

Make it easier to explore data with highlights.

Computing diversity measures on tripartite graphs. This package first implements a parametrized family of such diversity measures which apply on probability distributions. Sometimes called ‘True Diversity’, this family contains famous measures such as the richness, the Shannon entropy, the Herfindahl-Hirschman index, and the Berger-Parker index. Second, the package allows to apply these measures on probability distributions resulting from random walks between the levels of tripartite graphs. By defining an initial distribution at a given level of the graph and a path to follow between the three levels, the probability of the walker’s position within the final level is then computed, thus providing a particular instance of diversity to measure.

A pair of functions for renaming and encoding data frames using external crosswalk files. It is especially useful when constructing master data sets from multiple smaller data sets that do not name or encode variables consistently across files. Based on similar commands in ‘Stata’.

Fits meta-CART by integrating classification and regression trees (CART) into meta-analysis. Meta-CART is a flexible approach to identify interaction effects between moderators in meta-analysis. The methods are described in Dusseldorp et al. (2014) <doi:10.1037/hea0000018> and Li et al. (2017) <doi:10.1111/bmsp.12088>.

**22**
*Sunday*
Oct 2017

Posted Books

in
**22**
*Sunday*
Oct 2017

Posted Distilled News

in**Artificial Intelligence as a Service – AIaaS**

Suddenly, artificial intelligence is everywhere. Are you AI ready if not then be ready to be read in history books. Are we not missing the fact that artificial intelligence is about the people, not the machines. Technology and non technology companies are now investing and brining out the real and materialistic values of Artificial Intelligence to the real world. Its almost after a frustrating and hard work of decade AI has started delivering values. Using the contemporary view of computing exemplified by recent models and results from non-uniform complexity theory has proven the fact. Investment in artificial intelligence is growing fast. Tech giants like Google, Microsoft, Apple and Baidu known for their dominance in digital technologies globally are spending couple of tens of billions united state dollars on AI with 90 percent of this spent on R&D and deployment, and 10 percent on AI acquisitions. It takes money to make money and right now a lot of that money is going into the development of artificial intelligence. Any intelligence level surpassing the human intelligence is called the superintelligence level which is still 50 years plus ahead.

**Top 10 Machine Learning Algorithms for Beginners**

1. Linear Regression

2. Logistic Regression

3. CART

4. Naïve Bayes

5. KNN

6. Apriori

7. K-means

8. PCA

9. Bagging with Random Forests

10. Boosting with AdaBoost

2. Logistic Regression

3. CART

4. Naïve Bayes

5. KNN

6. Apriori

7. K-means

8. PCA

9. Bagging with Random Forests

10. Boosting with AdaBoost

**5 Free Resources for Furthering Your Understanding of Deep Learning**

This post includes 5 specific video-based options for furthering your understanding of neural networks and deep learning, collectively consisting of many, many hours of insights.

**Practical Machine Learning with R and Python – Part 3**

In this post ‘Practical Machine Learning with R and Python – Part 3’, I discuss ‘Feature Selection’ methods. This post is a continuation of my 2 earlier posts 1.Practical Machine Learning with R and Python – Part 1 2.Practical Machine Learning with R and Python – Part 2 While applying Machine Learning techniques, the data set will usually include a large number of predictors for a target variable. It is quite likely, that not all the predictors or feature variables will have an impact on the output. Hence it is becomes necessary to choose only those features which influence the output variable thus simplifying to a reduced feature set on which to train the ML model on.

Learn container orchestration with Kubernetes, featuring easy-to-use recipes for Kubernetes installation, API access, monitoring, troubleshooting, and more.

**22**
*Sunday*
Oct 2017

Posted Magister Dixit

in“Improvements in technology and big data trends have given rise to improvements in machine learning. The sheer volume of data is growing exponentially, and companies are looking for faster speeds and real-time analytics. Cognitive computing combines machine learning and artificial intelligence to go beyond data mining and provide actionable insights.” Gil Allouche ( January 9, 2015 )