Reinforcement Learning Guide: Solving the Multi-Armed Bandit Problem from Scratch in Python

Do you have a favorite coffee place in town? When you think of having a coffee, you might just go to this place as you´re almost sure that you will get the best coffee. But this means you´re missing out on the coffee served by this place´s cross-town competitor. And if you try out all the coffee places one by one, the probability of tasting the worse coffee of your life would be pretty high! But then again, there´s a chance you´ll find an even better coffee brewer. But what does all of this have to do with reinforcement learning?

10 Mind-Blowing TED Talks on Artificial Intelligence Every Data Scientist & Business Leader Must Watch

• The Incredible Inventions of Intuitive AI – Maurice Conti
• How Algorithms Shape our World – Kevin Slavin
• What Happens When our Computers Get Smarter than We Are? – Nick Bostrom
• Can we Build AI without Losing Control Over it? – Sam Harris
• How a Driverless Car Sees the Road – Chris Urmson
• How We´re Teaching Computers to Understand Pictures – Fei-Fei Li
• How Computers Learn to Recognize Objects Instantly – Joseph Redmon
• The Jobs We´ll Lose to Machines – Anthony Goldbloom
• How AI can Enhance our Memory, Work and Social Lives – Tom Gruber
• How AI can Compose a Personalized Soundtrack to your Life – Pierre Barreau

Data Science vs AI: Get to the Fundamentals

Deriving meaningful information out of heap of data is the minimal requirement for any establishment today for its survival & sustenance. There are many terminologies and buzz words related to this area that blurs the meaning leaving people confused, such as Bigdata, Data Ware house (DWH), BI analytics, AI (Artificial Intelligence), Data Science, Machine Learning, Advance Analytics, Deep Learning, Cognitive, Predictive modeling etc. to name a few. I have seen institutions would have team using machine learning to build classifiers but they call the same as AI team, Business Analysts would run some diagnostic analysis using Tableau but are called Data Scientists, sometimes we write conventional code or may be use RPA tool such as blue prism in automating certain portion of business process (e.g. take data from one app, paste to a file , format the same before sending as expense report) and we might unintentionally call that as AI eco-system and so on. College graduates, job seekers (fresh or lateral), business executives and technologists must make good effort to understand the concept behind various data management subject area (AI, ML, DS, Deep Learning, Cognitive Computing, Statistics, BI, DWH etc.), associated roles (such as data engineers, business analysts, data scientists, ML engineer, Data Modelers, data administrators etc.) and subsequently plan to learn and apply else Industries might lose revenue, CoE & practices blur and frustration creep in between expectation vs reality. We, however, haven´t gone that far in data maturity area (Exploration, diagnostic, prediction & prescriptive), hence important to clarify and understand various data subject area before its late. This article is an attempt to put clarity around these data subject area involving AI, Data Science and related terms to help graduates, data practitioners, business executives and others to develop career, establish practices, community and competency in Data Science area.

Conversational UI is our Future

Conversational user interface (UI) is changing the way that we interact. Intelligent assistants, chatbots and voice-enabled devices, like Amazon Alexa and Google Home, offer a new, natural, and intuitive human-machine interaction and open up a whole new world for us as humans. Chatbots and voicebots ease, speed up, and improve daily tasks. They increase our efficiency and compared to humans, they are also very cost effective for the businesses employing them. This article will address the concept of conversational UIs by initially exploring what they are, how they evolved, what they offer. The article provides an introduction to the conversational world. We will take a look at how UI has developed over the years and the difference between voice control, chatbots, virtual assistants, and conversational solutions.

Working with Data Feeds

As companies become more data-driven there is often a proliferation of data from both internal sources as well as third parties being consumed. Rarely have I seen firms try and centralise where datasets are stored. Instead, data is often copied onto infrastructure for individual teams and departments. This allows teams to not disrupt others with their work as well as avoid disruption from other teams. Data sources are often refreshed in batches ranging from every few minutes to monthly updates. The file formats, compression schemes and encryption systems used to proliferate these datasets can vary greatly. There is no one single tool I use for collection and analysis of new datasets. I do my best to pick tools that help me avoid writing a lot of bespoke code while taking advantage of the hardware available on any one system I may be using. In this guide I’ll walk through a exercise in consuming, transforming and analysing a data dump of the English language version of Wikipedia.

Where am I?

The here package is pretty simple ( only 3 functions), but I cannot remember how to use it to navigate folders, so this is my aide-memoire. It might be useful for others too. Here finds the root of your current folder / working directory. If you use Projects in RStudio, that will usually be the root of your project folder. If not, you can use set_here() to create a small file which will set the root location.

The Difference Between Link Functions and Data Transformations

Generalized linear models – and generalized linear mixed models – are called generalized linear because they connect a model´s outcome to its predictors in a linear way. The function used to make this connection is called a link function. Link functions sounds like an exotic term, but they´re actually much simpler than they sound.

Price Elasticity: Data Understanding and Data Exploration First Of All!

An experiment about modeling price elasticity as an example and, after analyzing the model with residual plots, it turned out there’s a problem after the 1st of September in the test data set …

What’s New in Deep Learning Research: Reducing Bias and Discrimination in Machine Learning Models…

What’s New in Deep Learning Research: Reducing Bias and Discrimination in Machine Learning Models with AI Fairness 360

Convolutional Neural Networks for Beginners: Practical Guide with Python and Keras

At this point, we are ready to deal with another type of neural networks, the so-called convolutional neuronal networks, widely used in computer vision tasks. These networks are composed of an input layer, an output layer and several hidden layers, some of which are convolutional, hence its name.

The secrets behind Reinforcement Learning

Bots that play Dota2, AI that beat the best Go players in the world, computers that excel at Doom. What’s going on? Is there a reason why the AI community has been so busy playing games? Let me put it that way. If you want a robot to learn how to walk what do you do? You build one, program it and release it on the streets of New York? Of course not. You build a simulation, a game, and you use that virtual space to teach it how to move around it. Zero cost, zero risks. That’s why games are so useful in research areas. But how do you teach it to walk? The answer is the topic of today’s article and is probably the most exciting field of Machine learning at the time:

Machine Learning — Text Processing

Text Processing is one of the most common task in many ML applications. Below are some examples of such applications.

Naive Bayes Classifier: A Geometric Analysis of the Naivete. Part 1

The curse of dimensionality is the bane of all classification problems. What is the curse of dimensionality? As the the number of features (dimensions) increase linearly, the amount of training data required for classification increases exponentially. If the classification is determined by a single feature we need a-priori classification data over a range of values for this feature, so we can predict the class of a new data point. For a feature x with 100 possible values, the required training data is of order O(100). But if there is a second feature y as well that is needed to determine the class, and y has 50 possible values, then we will need training data of order O(5000) – i.e. over the grid of possible values for the pair ‘x, y’. Thus the measure of the required data is the volume of the feature space and it increases exponentially as more features are added.

Spotify App review mining with R + Google Cloud Machine Learning

This article describes how to export data from iTunes with R and itunesr (by Abdul Majed Raja) followed by visualized ratings and reviews. It also covers how to use googleLanguageR for translating reviews by performing language processing with Google Cloud Machine Learning API before conducting basic sentiment analysis.