Monday, March 19, 2018

Google Machine Learning Crash Course


Machine Learning Crash Course  |  Google Developers

"A self-study guide for aspiring machine learning practitioners
Machine Learning Crash Course features a series of lessons with video lectures, real-world case studies, and hands-on practice exercises."

Artificial intelligence development course offered for free by Google - Science & Tech - The Jakarta Post

tool: PWA Builder, App Image Generator

Interesting tools:

Take a URL of running web site, download pages, and package as PWA that can run disconnected!

" takes data from your site and uses that to generate cross-platform Progressive Web Apps"

Take a single image, and create set of required logo images and thumbnails for 
Windows 10, Android, iOS etc... Very useful!

"Quickly and easily create app icons for various platforms in the right size and format"

link from .NET Rocks! vNext

Saturday, March 17, 2018

Markdown Editor for Visual Studio

Markdown Editor - Visual Studio Marketplace
by Mads Kristensen

Preview window

or as a par of:
Web Essentials 2017 - Visual Studio Marketplace

for previous versions of Visual Studio
Web Essentials for Visual Studio

Markdown editing with Visual Studio Code
"VS Code supports Markdown files out of the box. You just start writing Markdown text, save the file with the .md extension and then you can toggle the visualization of the editor between the code and the preview of the Markdown file; obviously, you can also open an existing Markdown file and start working with it. To switch between views, press Ctrl+Shift+V in the editor. You can view the preview side-by-side (Ctrl+K V) with the file you are editing and see changes reflected in real-time as you edit."

Friday, March 16, 2018

Graphs for Machine Learning

Graphs as the front end for machine learning - O'Reilly Media
The O’Reilly Data Show Podcast: Leo Meyerovich on building large-scale, interactive applications that enable visual investigations.



Thursday, March 15, 2018

IoT: Raspberry Pi 3 Model B+

Raspberry Pi 3 Model B+ on sale now at $35 - Raspberry Pi

"Raspberry Pi 3 Model B+ is now on sale for $35 (the same price as the existing Raspberry Pi 3 Model B), featuring:
  • A 1.4GHz 64-bit quad-core ARM Cortex-A53 CPU
  • Dual-band 802.11ac wireless LAN and Bluetooth 4.2
  • Faster Ethernet (Gigabit Ethernet over USB 2.0)
  • Power-over-Ethernet support (with separate PoE HAT)
  • Improved PXE network and USB mass-storage booting
  • Improved thermal management
Alongside a 200MHz increase in peak CPU clock frequency, we have roughly three times the wired and wireless network throughput, and the ability to sustain high performance for much longer periods."

100,000 Events Per Second on Azure Functions

Processing 100,000 Events Per Second on Azure Functions | Azure App Service Team Blog

link from:
Philly Azure March Meeting | Philly Azure (Malvern, PA) | Meetup

Wednesday, March 14, 2018

pi day: Calculate Pi in Azure Notebook

Interesting usage of iPython-like Notebooks for a web article:

This Pi Day, Calculate the Value of Pi for Yourself | WIRED
Calculating area of half-cicrle of radius r=1 as sum of vertical slices
The surface area of circle is r**2 * pi = pi  for r==1
Area of a slice is y * dx
x**2 + y**2 = r**2 = 1  =>  y = sqrt(1-x**2)

def calcPiSlices(n):
    dx = 2. / n # step size
    area = 0 # sum of all areas of half-circle
    for x in numpy.arange(-1, 1, dx):
        area = area + math.sqrt(1-x**2) * dx # area of a vertical slice
    return 2*area # *2 to get area of whole circle = pi

Here is the calculation in "real" Jupyter Notebook on 

Pi - Wikipedia  |  5 Ways to Calculate Pi - wikiHow

faster: Nilakantha series method
π = 3 + 4/(2*3*4) - 4/(4*5*6) + 4/(6*7*8) - 4/(8*9*10) + 4/(10*11*12) - 4/(12*13*14) ... 

def calcPiNilakantha(n):
    pi = 3.0
    for x in range(2, n, 4):
        pi += (4 / (x*(x+1)*(x+2))) - (4 / ((x+2)*(x+3)*(x+4)))
    return pi

Azure GeoAI by Microsoft & Esri

Microsoft and Esri launch Geospatial AI on Azure | Blog | Microsoft Azure

"Microsoft and Esri will be offering the GeoAI Data Science Virtual Machine (DSVM) as part of our Data Science Virtual Machine/Deep Learning Virtual Machine family of products on Azure."

Tuesday, March 13, 2018 cloud architecture diagrams tools

Online drawing tools, often used for (AWS) cloud diagrams.  it is free (for single user), and open source JGraph @ GitHub

Integrations – : works with Google Docs (GSuite)
"highest user rated productivity tool in the Chrome Web Store" – Online Diagramming

AWS diagram

AWS Simple Icons

other drawing tools:

Lucidchart Libraries »
AWS Marketplace listing »

Cacoo Store
Design Pattern Templates

Creately Templates Diagram Editor

Cloudcraft smart AWS diagram editor

DC Solution Factory
Visual Architecture and Deployment

Download Microsoft Azure, Cloud and Enterprise Symbol / Icon Set
 - Visio stencil, PowerPoint, PNG, SVG from Official Microsoft Download Center

Icons and other assets for architectual diagrams | Microsoft Docs

Google Cloud Platform Blog: Google Cloud Platform icons and sample architectural diagrams, for your designing pleasure

free ebook: Getting Started with AI

"Getting Started with Artificial Intelligence:
A Practical Guide to Building Enterprise Applications
is full of case studies, implementation examples, and a survey of the current landscape. It's an ideal starting point for your journey into enterprise AI. And it's free courtesy of IBM."

AI class: How Google does Machine Learning @ Coursera

How Google does Machine Learning | Coursera

"About this course: What is machine learning, and what kinds of problems can it solve? Google thinks about machine learning slightly differently -- of being about logic, rather than just data. We talk about why such a framing is useful when thinking about building a pipeline of machine learning models. Then, we discuss the five phases of converting a candidate use case to be driven by machine learning, and consider why it is important the phases not be skipped. We end with a recognition of the biases that machine learning can amplify and how to recognize this.

... learning how to build and operationalize TensorFlow models."

some other ML classes:
  • Machine Learning with Andrew Ng coursera and Deep Learning Specialization coursera
  • site
  • Machine Learning with John W. Paisley edx
  • Machine Learning Columbia University edx
links from:

Monday, March 12, 2018

Microsoft Translation AI: DNN vs. statistical

Machine translation of human languages has improved significantly in last 10 years, but it is still far from perfect. Here are very good articles explaining how this works now,
and interesting web site that compares various methods of translation for select languages.

Machine Translation - Microsoft Translator

Both SMT (Statistical Machine Translation) and
Deep Neural Network) translation technologies
have two elements in common:
  • Both require large amounts of pre-human translated content (up to millions of translated sentences) to train the systems.
  • Neither act as bilingual dictionaries, translating words based on a list of potential translations, but translate based on the context of the word that is used in a sentence."
Try and Compare - Microsoft Translator
Artificial Intelligence, powered by neural networks:
used to be about 20 languages, now only Chinese; site sometimes works

Classic "statistical" translator: about 60 languages; site works, and this is same as API

marketing pages:

Try Microsoft Translator for free:
Bing, Live conversations, Office 365 Word,

Translator Text API - Auto Translate | Microsoft Azure

API documentation:

sample code calling Microsoft Translation API (classic):

Google paper:

Friday, March 09, 2018

Blog to Podcast, via AWS Polly cloud service

to skip "Alexa, read me this blog post"

Amazon launches a Polly WordPress plugin that turns blog posts into audio, including podcasts - AWS in the News

Amazon Polly for WordPress — WordPress Plugins

"Amazon Polly is a service that turns text into lifelike speech. With dozens of voices across a variety of languages, you can select the ideal voice and build engaging speech-enabled applications that work in many different countries. The Amazon Polly plugin for WordPress is a sample application that shows how WordPress creators can easily add Text-to-Speech capabilities to written content with Amazon Polly. You can generate an audio feed for text-based content and insert it into an embedded player to increase the accessibility of your WordPress site. The sample code also enables you to publish podcasts directly from your site and make them available for listeners in the form of podcasts."

Amazon launches a Polly WordPress plugin that turns blog posts into audio, including podcasts | TechCrunch

Amazon Polly Pricing

"The Amazon Polly free tier includes 5 million characters per month for speech or Speech Marks requests, for the first 12 months, starting from your first request for speech.
$4.00 per 1 million characters for speech (~23 hours, 8 min)

Microsoft Azure "Cognitive Service" / Bing has a similar service / API, at least it used to have it.
Speech-to-text is still there. Current list of service does not include text-to-speech.
Similar API is available in Windows, but with smaller number of "Agent" voices.

Text to Speech API of Microsoft Speech Service | Microsoft Docs

Cognitive Service Try experience | Microsoft Azure

GCP also includes only speech-to-text, , no text-to-speech, only provides "reader" Android app.

Cloud Natural Language  |  Google Cloud Platform

Related, Amazon also started offering translation service with limited set of languages, but extensible

Amazon Translate – Neural Machine Translation - AWS

A quick look at Machine Translation with Amazon Translate

Thursday, March 08, 2018

IoT startup by founder of Netscape: CommandScape

CommandScape - YouTube

Billionaire Netscape founder Jim Clark is back with a new, self-funded startup | TechCrunch

"CommandScape, a home-automation and building-control system that controls air conditioning, motorized shades, surveillance, lighting, fire alarms and other functions from a highly secure online platform that has cybersecurity at its core.

electronics journal CE ProCommandScape’s system is completely wired, delivering data and power to some products over a building’s existing AC wiring.

...passwords and usernames aren’t needed... Users are instead authenticated via their iPhones (with a systems administrator, or a parent, able to revoke permissions).

It could be an enormous company . . . in a slow-moving [$100 billion] market where technology is 15 years old,” Clark, now 73, told USA Today"

CommandScape : "Control and automate security, access control, lighting, HVAC, video surveillance, and more. All from one sleek and cybersecure integrated app."

Meet the Team"Jim Clark is the founder of CommandScape TM, a fully integrated cybersecure operating and control system for commercial buildings and larger homes. Jim is a prolific entrepreneur and founder of Netscape, Silicon Graphics, WebMD, MyCFO & Shutterfly."

Our Story"CommandScapeTM integrates its cybersecure alarm panel, climate, access and lighting controls, video surveillance and connected devices into both residential and commercial installations of any size."

Technology: TLS, iOS, AWS, GoLang, Kubernetes

Wednesday, March 07, 2018

Windows Developer Day: ML / AI, IoT

Windows Developer Day – March 7, 2018 - Windows app development
Focus of next update of Windows 10:
  1. Modern Application Experiences
  2. Intelligent Edge & Windows AI Platform
  3. Windows Internet of Things
Desktop Bridge: Win32 + UWP

WebView control: Edge instead of IE 11; open source; 
running in a separate process, more robust, WebRTC
for WinForms, WPF and UWP

"Containers for (Windows) apps" : .MSIX = .appx + .MSI
(for all Win32, WinForms, WPF, UWP apps)
open sourced MSIX format
Microsoft/msix-packaging: MSIX Packaging SDK @ GitHub
"Advanced Installer" tool, VS integration

Three things you need to know about Windows Developer Day - Windows Developer BlogWindows Developer Blog

Windows 10’s next major update will include Windows ML, a new AI platform - The Verge
"(New AI platform) Windows ML will enable developers to create more powerful apps for consumers running Windows 10. Developers will be able to import existing learning models from different AI platforms and run them locally on PCs and devices running Windows 10"

Google Cloud Spanner (vs. Azure Cosmos DB vs. Amazon DynamoDB)

GCP Spanner is a very unusual tool: from outside it behaves as relational database,
but internals are more similar to NoSQL databases, scaling horizontally, no downtime.
From outside it behaves as a single server, but in fact it is globally distributed.
And it is production grade, tested for many years on Google scale.
In fact GCP itself runs on Spanner.

Cloud Spanner | Automatic Sharding with Transactional Consistency at Scale  |  Google Cloud Platform
"Cloud Spanner is the only enterprise-grade, globally-distributed, and strongly consistent database service built for the cloud specifically to combine the benefits of relational database structure with non-relational horizontal scale. This combination delivers high-performance transactions and strong consistency across rows, regions, and continents with an industry-leading 99.999% availability SLA, no planned downtime, and enterprise-grade security."

Cloud Spanner 101: Google's mission-critical relational database (Google Cloud Next '17) - YouTube

Cloud Spanner 201: getting the most out of Cloud Spanner (Google Cloud Next '17) - YouTube

Spanner, the Google Database That Mastered Time, Is Now Open to Everyone | WIRED
"...they equipped Google's data centers with a series of GPS receivers and atomic clocks...
...This gives you faster-than-light coordination between two places," says Peter Mattis, a former Google engineer who founded CockroachDB, a startup working to build an open source version of Spanner."

Google Research Publication: Spanner (original paper explaining how it works)

Cloud Spanner Documentation  |  Cloud Spanner  |  Google Cloud Platform

Discover on Google Cloud Spanner – DeepQ Research Engineering Blog – Medium

Google's Cloud Spanner: how does it stack up? | ZDNet

"...Spanner is a relational database, geared to operational OLTP (online transactional processing) workloads, with full ACID (atomicity, consistency, isolation and durability) functionality. Spanner's not a simple scale-up relational database service -- that's where Google Cloud SQL comes in. Spanner is not a data warehouse; Google BigQuery is designed to handle those workloads. And it's not a NoSQL database, either, as BigTable is Google's offering there.

So Spanner contrasts strongly with Amazon's DynamoDB, which is a NoSQL database employing so called "eventual consistency," and Microsoft's Cosmos DB, also a NoSQL database, and one which is configurable along a full spectrum of consistency models, ranging from an ACID model on one end to eventual consistency on the other, and two more consistency models in between."

Cloud Database Wars: Google Spanner vs. Microsoft CosmosDB | Simplilearn

Google Cloud Spanner vs. Microsoft Azure Cosmos DB Comparison

Practical Tradeoffs in Google Cloud Spanner, Azure Cosmos DB and YugaByte DB

A Chart Comparing Azure Cosmos DB Amazon DynamoDB and Google Cloud Spanner
GCP Spanner public features have changed since this chart was created...

Monday, March 05, 2018

How IPython, Jupyter Notebooks work

Jupyter Notebooks are favorite GUI for Data Science and Machine Learning documentation.
They provide and interactive web environment for editing text (in simple Markdown syntax)
and mix with program code that can run interactively, with data, on an connected server.

It started as an interface for Python, called IPython, and evolved to support related languages Julia and R, expanding and renaming project to Jupyter. Now there are 50+ languages supported.
The system is running in 3 places: JavaScript in web browser, that communicates
with web server application that is managing state and is communicating
with "Kernel" process that runs interactive code.

How IPython and Jupyter Notebook work — Jupyter Documentation 4.1.1 alpha documentation
"The Notebook frontend does something extra. In addition to running your code, it stores code and output, together with markdown notes, in an editable document called a notebook. When you save it, this is sent from your browser to the notebook server, which saves it on disk as a JSON file with a .ipynb extension.

How to Use Jupyter Notebooks in Azure Machine Learning Workbench | Microsoft Docs

notebook architecture

Jupyter kernels · jupyter/jupyter Wiki

dsblank/simple_kernel: A small, simple kernel example for IPython

excellent interview:
Azure Notebooks | Azure Friday | Channel 9

Microsoft Azure Notebooks

book as Jupyter Notebook @ Azure:
PythonDataScienceHandbook (jakevdp) - Microsoft Azure Notebooks
book text:
Python Data Science Handbook | Python Data Science Handbook

notebooks from another book:
amueller/introduction_to_ml_with_python: Notebooks and code for the book "Introduction to Machine Learning with Python" @ GitHub

Jupyter Notebook conference & training: JupyterCon

very interesting recorded presentations from previous conferences!
JupyterCon - YouTube - YouTube 

Project Jupyter: From interactive Python to open science - Fernando Perez - YouTube

free class lectures:
Data 8: Foundations of Data Science | Division of Data Sciences @
Introduction · Computational and Inferential Thinking @ GitBooks

Computational Publishing with Jupyter by O'Reilly Media

odewahn/computational-publishing @ GitHub
= Content in Git + Docker file + Runtime

Oriole Online Tutorials
"...complete integration of video synchronized with the flow of the text, as well as the ability to execute the code: this is probably as close as we can get to learning side-by-side"

Sunday, March 04, 2018

iCloud @ AWS S3, GCP

Apple may no longer be using Microsoft's Azure | ZDNet

"The (iCloud) service is agnostic about what is being stored and handles all file content the same way, as a collection of bytes," according to Apple's 2018 security guide. "Each file is broken into chunks and encrypted by iCloud using AES-128 and a key derived from each chunk's contents that utilizes SHA-256. The keys and the file's metadata are stored by Apple in the user's iCloud account.The encrypted chunks of the file are stored, without any user-identifying information, using third-party storage services, such as S3 and Google Cloud Platform."

free cloud: Azure for College/University Students

The offer is similar to regular trial offer, with some services free for 12 months and some "always free". The difference is that instead of credit card the user needs to have email address from accredited educational institution.

Azure for Students—Azure FAQ | Microsoft Azure

"Azure for Students is available only to students who meet the following requirements. You must affirm that you are age 18 or older and attend an accredited, degree-granting four-year educational institution where you’re a full-time student or faculty in a STEM-related field. You must verify your academic status through your organization’s email address (or Shibboleth, if supported by your organization). This offer is not available for use in a massive open online course (MOOC) or in other professional trainings from for-profit organizations."

Saturday, March 03, 2018

cloud services: Azure vs AWS

Excellent comparison of AWS and Azure services

Azure and AWS services compared - multicloud | Microsoft Docs

Azure for AWS Professionals | Microsoft Docs
Comparison of structure and ownership of AWS accounts and Azure subscriptions

Visual Studio (Code) + Anaconda

For doing Data Science, Anaconda tools for Python are the most popular platform.
In recent versions of Visual Studio 2017 Anaconda started appearing as a (large) option.
Now, there is also option to get VS Code with Anaconda.

VS Code in Anaconda Distribution 5.1 | Anaconda

Anaconda and Microsoft Partner to Deliver Python-Powered Machine Learning | Business Wire

Friday, March 02, 2018

IoT security: MIT’s low power encryption chip

more custom chips...

MIT’s low power encryption chip could make IoT devices more secure

"...encryption... is energy-intensive. MIT is working on a new chip, however, to perform this sort of public-key encryption that only uses 1/400 as much power as a software solution would. In addition, the chip uses about 1/10 as much memory and executes processes 500 times as fast."

Could specialized HW like this also help adoption of block-chain technology?

IEEE Computing Edge

"Edge Computing" is a "new hot" technology.
Many tools and techniques optimized for "cloud" are as well useful on remote locations,
and can be managed in unified way, in particular useful for IoT and ML.

Great examples:

  • Microsoft Azure Stack (HW+SW) and Azure IoT Edge (SW), as well as 
  • Amazon Snowball Edge (HW+SW), as well as GreenGrass (SW)

IEEE (free) publication covering this:

ComputingEdge - Your View into the Entire World of Leading-Edge Technology • IEEE Computer Society

"ComputingEdge curates the hot technology knowledge from our 13 leading technology publications, plus adds unique original content, and makes it available to you all in a single experience"

Thursday, March 01, 2018

Free cloud for startups: Azure, AWS, GCP, IBM, Oracle...

Microsoft has recently replaced "Biz Spark" program
with a new  "Microsoft for Startups", focused on Azure.

Program Benefits | Microsoft for Startups
"Up to $120K of free Azure cloud credits*"

link from:
Microsoft Cloud Show - New Microsoft Support for Startups along with Azure and SharePoint News

GCP and AWS have similar:
$3K to $100K for one year
up to $100K in 2 years

All cloud service providers have excellent "getting started" free options for individuals:

AWS vs. Azure vs. Google Cloud: Which free tier is best? | InfoWorld

AWS Free Tier

AWS Free Tier includes offers that expire 12 months following sign up
+ others that never expire.
Create your Azure free account today | Microsoft Azure

$200 credit: to explore services for 30 days
+12 months: of popular free services+ Always free: 25+ services

GCP Free Tier - Free Extended Trials and Always Free | Google Cloud Platform

12 Months: $300 free credit to get started with any GCP product.
+ Always Free : Free usage limits on participating products for eligible customers, during and after the free trial. Offer is subject to change.

Sign up for IBM Cloud
IBM Cloud pricing | IBM Cloud

Guaranteed free development with Lite plans
Develop worry-free and at no cost with cap based Lite plan services for as long as you like.

Get Started with Oracle Cloud for Free $300 in free credits

"Dynos are smart, lightweight containers built for modern languages and developer productivity"
free: 512 MB RAM │ 1 web/1 worker