via prostheticknowledge.tumblr.com

Calvin & Hobbes: the big bang

Project by Gabriel de Laubier turns a comic strip into an interactive 3D scene:

3D rendition of a Calvin & Hobbes comic strip (orginal strip by Bill Watterson). Backface culling and flat-shading are used to give life to the cartoon drawings. You can see it in realtime on Sketchfab 

View for yourself here

via flowingdata.com

I typically don’t care much about code releases, but every interactive chart I make these days uses Mike Bostock’s JavaScript library D3.js. It get more fun the more you use it. And it just got better with a 4.0 release.

D3 4.0 is modular. Instead of one library, D3 is now many small libraries that are designed to work together. You can pick and choose which parts to use as you see fit. Each library is maintained in its own repository, allowing decentralized ownership and independent release cycles. The default bundle combines about thirty of these microlibraries.

More importantly:

Small files are nice, but modularity is also about making D3 more fun. Microlibraries are easier to understand, develop and test. They make it easier for new people to get involved and contribute. They reduce the distinction between a “core module” and a “plugin”, and increase the pace of development in D3 features.

Nice.

Tags:d3js, Mike Bostock

via prostheticknowledge.tumblr.com

ISHU

Brand of scarf with anti-paparazzi properties to sabotage flash photography popular with celebrities:

In a world where the choice to remain anonymous is no longer a choice, Access All Brands launches the first ever consumer product precision-engineered to claim back your right to privacy, in a sophisticated and sexy manner.

Presenting  The  ISHU, a world premier fashion scarf that allows the wearer to  influence  flash  photography!   Saif  Siddiqui,  founder of Access All Brands  and  the  force behind the creation of this #InvisibilityCloak  was adamant that a stylish solution be available to that select group of people who  want  to  control  unwanted  pictures  of them being taken with mobile devices, which inevitably end up plastered across social media. Simply put, The ISHU gives you control over your image and as such, the power to decide which  pictures  of  you  end  up on Snapchat, Instagram, Facebook.

You can find out more about the product here

june1777 posted a photo:

june1777 posted a photo:

via www.thecitizenscampaign.org

We would like to welcome our newest team member, Joel Rosa! Joel will serve as our Civic Challenge Facilitator in charge of promoting the Challenge and to help motivate citizens build solutions-focused communities in their own cities and towns. Joel is an active and dedicated member of his community. He is a member of the Perth Amboy Knights of Columbus, Vice-President of the Central Jersey Latino Chamber of Commerce, and a board member of La Casa de Educaccion y Cultura Latina, whose mission is to advance the culture and education of people from Latin America residing in the United States. Joel is also a board member of Hugs Towards Corporation, a non-profit dedicated to providing educational supports for children of families in need. Joel received his Bachelor of Science degree from Montclair State University. Reach out to Joel to see how you can get involved by emailing him at joel@thecitizenscampaign.org

via vanishingnewyork.blogspot.com

via prostheticknowledge.tumblr.com

RAID: A Relation-Augmented Image Descriptor

Computer Vision research from the Smart Geometry Processing Group can retrieve images based on visual descriptions created with doodles:

As humans, we regularly interpret scenes based on how objects are related, rather than based on the objects themselves. For example, we see a person riding an object X or a plank bridging two objects. Current methods provide limited support to search for content based on such relations. We present RAID, a relation-augmented image descriptor that supports queries based on inter-region relations. The key idea of our descriptor is to encode region-to-region relations as the spatial distribution of point-to-region relationships between two image regions. RAID allows sketch-based retrieval and requires minimal training data, thus making it suited even for querying uncommon relations. We evaluate the proposed descriptor by querying into large image databases and successfully extract nontrivial images demonstrating complex inter-region relations, which are easily missed or erroneously classified by existing methods. We assess the robustness of RAID on multiple datasets even when the region segmentation is computed automatically or very noisy.

More Here

via prostheticknowledge.tumblr.com

Prisma

Russian iOS photo app does a pretty good job in employing neural net style transfer to your smartphone photos:

Prisma transforms your photos into artworks using the styles of famous artists: Van Gogh, Picasso, Levitan, as well as world famous ornaments and patterns. A unique combination of neural networks and artificial intelligence helps you turn memorable moments into timeless art. 

You can find out more here

Image credits: ellineya, baustana, eliztimofeeva and elizabet_myr

june1777 posted a photo:

via www.boozyburbs.com

Hackensack’s first microbrewery, The Alementary Brewing Company, will release a Hackensack commemorative beer called the Hackensack Lager in their tasting room this Sunday, July 17th at 12 noon. This brewery exclusive beer is a light lager that should have “broad appeal” for all types of beer drinkers to go along with their existing well-received brews.drinkers.

Since nothing says “summer” like beer and BBQ, Kimchi Smoke in Bergenfield is bringing their Texas-style barbecue to The Alementary to help celebrate this new beer.  Keeping with the Hackensack theme, Kimchi Smoke owner and chef Robert Austin Cho was a resident for over 10 years.

In addition to the Hackensack Lager, The Alementary will have their complete line-up of craft beers including the A-Game IPA, Mr. Stevens English Mild, and the Key Lime Gose.

The Alementary58 Voorhis Lane, Hackensack, NJ Website

The post Alementary Brewing in Hackensack to Launch Hackensack Lager appeared first on Boozy Burbs.

via prostheticknowledge.tumblr.com

Farmbot

Open source product is a robotic computer controlled home farming setup to grow your own food:

FarmBot Genesis is humanity’s first open-source CNC farming machine designed for at-home automated food production. 

More Here

via machinelearningmastery.com

It can be difficult to get started in deep learning.

Thankfully, a number of universities have opened up their deep learning course material for free, which can be a great jump-start when you are looking to better understand the foundations of deep learning.

In this post you will discover the deep learning courses that you can browse and work through to develop and cement your understanding of the field.

This is a long post that deep links into many videos. It is intended for you to bookmark, browse and jump into specific topics across courses rather than pick one course and complete it end-to-end.

Let’s get started.

Overview

We will take a quick look at the following 6 deep learning courses.

1. Deep Learning at Oxford
2. Deep Learning at Udacity by Google
3. Deep Learning Summer School at Montreal
4. Deep Learning for Natural Language Processing at Stanford
5. Convolutional Neural Networks for Visual Recognition at Stanford
6. Neural networks Class at Université de Sherbrooke

There is also a “Other Courses” section at the end to gather additional video courses that are not free, broken or smaller in scope and don’t neatly fit into this summary review.

Course Tips and How To Use This Post

There are a lot of courses and a lot of great free material out there.

My best advice is:

Do not pick a course and work through it end-to-end.

This is counter to what most people suggest.

Your impulse will be to “get serious” and pick “the best” course and work through all of the material. You will almost certainly fail.

The material is tough and you will need to take your time and get multiple different perspectives on each topic.

The very best way to really get into this material is to work through it topic by topic and draw from across all of the courses until you really understand a topic, before moving onto the next topic.

You do not need to understand all topics and you do not need to use a single source to understand a single topic.

Bookmark this page, then browse, sample and dip into the material you need, when you need it as you learn how to implement actual real deep learning models in code using a platform like Keras.

Get Started in Deep Learning With Python

Deep Learning gets state-of-the-art results and Python hosts the most powerful tools.
Get started now!

PDF Download and Email Course.

FREE 14-Day Mini-Course on 
Deep Learning With Python

 Download your PDF containing all 14 lessons.

Get your daily lesson via email with tips and tricks.

Deep Learning at Oxford

This is a machine learning course that focuses on deep learning taught at Oxford by Nando de Freitas.

I really like this course. I watched all of the videos on double time and took notes. It provides a good foundation in theory and covers modern deep learning topics such as LSTMs. Code examples are shown in Torch.

I noted that the syllabus differed from the actual video lectures available and the YouTube playlist listed the lectures out of order, so below is the list of 2015 video lectures in order.

• Deep Learning Lecture 1: Introduction
• Deep Learning Lecture 2: linear models
• Deep Learning Lecture 3: Maximum likelihood and information
• Deep Learning Lecture 4: Regularization, model complexity and data complexity (part 1)
• Deep Learning Lecture 5: Regularization, model complexity and data complexity (part 2)
• Deep Learning Lecture 6: Optimization
• Deep learning Lecture 7: Logistic regression, a Torch approach
• Deep Learning Lecture 8: Modular back-propagation, logistic regression and Torch
• Deep Learning Lecture 9: Neural networks and modular design in Torch
• Deep Learning Lecture 10: Convolutional Neural Networks
• Deep Learning Lecture 11: Max-margin learning, transfer and memory networks
• Deep Learning Lecture 12: Recurrent Neural Nets and LSTMs
• Deep Learning Lecture 13: Alex Graves on Hallucination with RNNs
• Deep Learning Lecture 14: Karol Gregor on Variational Autoencoders and Image Generation
• Deep Learning Lecture 15: Deep Reinforcement Learning – Policy search
• Deep Learning Lecture 16: Reinforcement learning and neuro-dynamic programming

The highlight for me was Alex Graves‘ talk on RNNs (Lecture 13). A smart guy doing great work. I was reading a lot of Alex’s papers at the time I watch this video so I may be biased.

Resources

• YouTube Playlist
• Course Summary and Materials
• Nando de Freitas homepage

Deep Learning at Udacity by Google

This is a mini course collaboration between Arpan Chakraborty from Udacity and Vincent Vanhoucke, a Principal Scientist at Google.

The course is free, hosted on Udacity and focuses on TensorFlow. It is a small piece of the broader Machine Learning Engineer Nanodegree by Google hosted on Udacity.

You must sign-up to Udacity, but once you sign-in you can access this course for free.

All course videos are on YouTube, but (intentionally) really hard to find with poor naming and linking. If anyone knows of a pirate playlist with all of the videos, please post it in the comments.

The course is divided into 4 lessons:

• Lesson 1: From Machine Learning to Deep Learning
• Lesson 2: Deep Neural Networks
• Lesson 3: Convolutional Neural Networks
• Lesson 4: Deep Models for Text and Sequences

The course is short but is broken up into many short videos and the Udacity interface is nice. Vincent seems to present in all of the videos I looked at (which is great) and videos are shown in the YouTube interface.

There is also a discussion form where you can ask and answer questions, driven by the slick discourse software.

My preference was to dip into videos that interested me rather than completing the whole course or doing any of the course work.

Resources

• Udacity Deep Learning Course
• Course Announcement on the Google Blog

Deep Learning Summer School at Montreal

A deep learning supper school was held in 2015 at the University of Montreal.

According to the website, the summer school was aimed at graduate students and industrial engineers and researchers who already have some basic knowledge of machine learning.

There were at least 30 talks (there are 30 videos) from notable researchers in the field of deep learning on a range of topics from introductory material to state of the art research.

Deep Learning Summer School at Montreal

These videos are are a real treasure trove. Take your time and pick your topics carefully. All videos are hosted on the VideoLectures.net site, which has an good enough interface, but not as clean as YouTube.

Many (all?) talks had PDF slides linked below the video, and more information is available from the schedule page on the official website.

Here’s the full list of lecture topics with links to the videos. I’ve tried to list related videos together (e.g. part 1, part 2).

• Introduction to Machine Learning
• Deep Learning: Theoretical Motivations
• Multilayer Neural Networks
• Training Deep Neural Networks
• Multilayer Neural Networks
• Deep Learning for Distribution Estimation
• Undirected Graphical Models
• Stacks of Restricted Boltzmann Machines
• On manifolds and autoencoders
• Visual features: From Fourier to Gabor
• Visual features II
• Convolutional Networks
• Learning to Compare
• NLP and Deep Learning 1: Human Language & Word Vectors
• NLP and Deep Learning 2: Compositional Deep Learning
• Seeing People with Deep Learning
• Deep Learning
• Deep Learning 2
• Speech Recognition and Deep Learning
• Tutorial on Neural Network Optimization Problems
• Deep Learning (hopefully faster)
• Adversarial Examples
• From Language Modelling to Machine Translation
• Deep NLP Recurrent Neural Networks
• Deep NLP Applications and Dynamic Memory Networks
• Memory, Reading, and Comprehension
• Smooth, Finite, and Convex Optimization Deep Learning Summer School
• Non Smooth, Non Finite, and Non Convex Optimization
• Variational Autoencoder and Extensions
• Deep Generative Models

Pick a topic and dive in. So good!

It looks like there will be a 2016 summer school and hopefully there will be videos.

Resources

• Deep Learning Summer School Official Website
• Deep Learning Summer School Videos

Deep Learning for Natural Language Processing at Stanford

This is a deep learning course focusing on natural language processing (NLP) taught by Richard Socher at Stanford.

An interesting note is that you can access PDF versions of student reports, work that might inspire you or give you ideas.

The YouTube playlist has poorly named files and some missing lectures. The 2016 videos are not all uploaded yet. Below is a list of the 2015 lectures and the links to the videos. Much easier for just jumping into a specific topic.

• Lecture 1: Intro to NLP and Deep Learning
• Lecture 2: Simple Word Vector representations: word2vec, GloVe
• Lecture 3: Advanced word vector representations: language models, softmax, single layer networks
• Lecture 4: Word Window Classification and Neural Networks
• Lecture 5: Project Advice, Neural Networks and Back-Prop (in full gory detail)
• Lecture 6: Practical tips: gradient checks, overfitting, regularization, activation functions, details
• Lecture 7: Recurrent neural networks for language modeling and other tasks
• Lecture 7 (8!?): GRUs and LSTMs for machine translation
• Lecture 9: Recursive neural networks for parsing
• Lecture 10: Recursive neural networks for different tasks (e.g. sentiment analysis)
• Lecture 11: Review Session for Midterm
• Lecture 13: Convolutional neural networks for sentence classification
• Lecture 15: Applications of DL to NLP
• Guest Lecture with Andrew Maas: Speech recognition
• Guest Lecture with Jason Weston: Memory Networks
• Guest Lecture with Elliot English: Efficient Implementations and GPUs

This is great material if you are into deep learning for NLP, an area where it really excels.

Resources

• CS224d: Deep Learning for Natural Language Processing Homepage
• Course Syllabus
• 2015 Course Video Play List
• 2016 Course Video Play List

Convolutional Neural Networks for Visual Recognition at Stanford

This course focuses on the use of deep learning for computer vision applications with convolutional neural networks.

It is another course taught at Stanford, this time by Andrej Karpathy and others.

Unfortunately, the course videos were taken down, but some clever people have found ways to put them back up in other places. See the playlists in the resources section below.

I regret to inform that we were forced to take down CS231n videos due to legal concerns. Only ¼ million views of society benefit served

— Andrej Karpathy (@karpathy) May 3, 2016

Another great course.

Below are the video lectures for the 2016 course, but I’m not sure how long the links will last. Leave a comment and let me know if you discover the links turned bad and I’ll fix them up.

• Lecture 1 Introduction and Historical Context
• Lecture 2 Data driven approach, kNN, Linear Classification 1
• Lecture 3 Linear Classification 2, Optimization
• Lecture 4 Backpropagation, Neural Networks 1
• Lecture 5 Neural Networks Part 2
• Lecture 6 Neural Networks Part 3 Intro to ConvNets
• Lecture 7 Convolutional Neural Networks
• Lecture 8 Localization and Detection
• Lecture 9 Visualization, Deep Dream, Neural Style, Adversarial Examples
• Lecture 10 Recurrent Neural Networks, Image Captioning, LSTM
• Lecture 11 ConvNets in practice
• Lecture 12 Deep Learning libraries
• Lecture 13 Segmentation, soft attention, spatial transformers
• Lecture 14 Videos and Unsupervised Learning
• Lecture 15 Invited Talk by Jeff Dean

Resources

• Convolutional Neural Networks for Visual Recognition Homepage
• Course Syllabus (and the 2015 Syllabus)
• Course Videos on Archive.org
• Course Videos on YouTube
• Course Sample Code

Neural Networks Class at Université de Sherbrooke

This is a course on neural networks taught by Hugo Larochelle at the University in Sherbrooke in Québec.

There is a ton of material. A ton.

The videos are one-on-one rather than lectures and there are many small videos for each topic rather than large one hour info dumps.

I think this might be a better format than the traditional lectures, but I’m not completely won over yet. A difficulty is there are 92 videos (!!!) to browse and it can be hard to find specific videos to watch.

The material is taught covering 10 main topics:

• Topic 1: Feedforward neural networks
• Topic 2: Training neural networks
• Topic 3: Conditional random fields
• Topic 4: Training Conditional random fields
• Topic 5: Restricted Boltzmann machine
• Topic 6: Autoencoders
• Topic 7: Deep learning
• Topic 8: Sparse coding
• Topic 9: Computer vision
• Topic 10: Natural language processing

My recommendation is to use the main course home page to browse the topics and then use those links into the specific videos. The YouTube playlist has far too many videos to browse and understand. The paradox of choice will kill you.

Resources

• Course Homepage
• YouTube Video Playlist

Other Courses

Below are some additional video courses that are either not free, difficult to access or smaller in scope.

• Deep Learning Course at CILVR Lab @ NYU (broken links?)
• Smaller Deep Learning Courses on Udemy
• Deep Learning at CMU
• Nvidia Self-Paced Courses for Deep Learning
• Neural Networks for Machine Learning at Coursera by the University of Toronto (awesome, but no longer free)
• Graduate Summer School: Deep Learning, Feature Learning, 2012

Do You Want To Get Started With Deep Learning?

You can develop and evaluate deep learning models in just a few lines of Python code. You need:

Deep Learning With Python

Take the next step with 14 self-study tutorials and
7 end-to-end projects.

Covers multi-layer perceptrons, convolutional neural networks, objection recognition and more.

Ideal for machine learning practitioners already familiar with the Python ecosystem.

Bring Deep Learning To Your Machine Learning Projects

Summary

In this post you have discover a number of world class video courses on deep learning covering, theory, computer vision, natural language processing and more.

Heed the advice at the top of this post.

Browse and dip into lectures by topic and do not try to take on a whole course. Learn one thing rather than try and learn everything.

Take your time, bookmark this page so you can come back, and have fun.

Do you know of some other video courses on deep learning that I have not listed? Let me know in the comments and I will update the list.

The post Deep Learning Courses appeared first on Machine Learning Mastery.

via prostheticknowledge.tumblr.com

Creative Applications of Deep Learning with TensorFlow

New online course from Kadenze put together by PK Mital will teach you how to use Google’s machine learning platform Tensorflow for creative projects:

This course introduces you to deep learning: the state-of-the-art approach to building artificial intelligence algorithms. We cover the basic components of deep learning, what it means, how it works, and develop code necessary to build various algorithms such as deep convolutional networks, variational autoencoders, generative adversarial networks, and recurrent neural networks. A major focus of this course will be to not only understand how to build the necessary components of these algorithms, but also how to apply them for exploring creative applications. We’ll see how to train a computer to recognize objects in an image and use this knowledge to drive new and interesting behaviors, from understanding the similarities and differences in large datasets and using them to self-organize, to understanding how to infinitely generate entirely new content or match the aesthetics or contents of another image. Deep learning offers enormous potential for creative applications and in this course we interrogate what’s possible. Through practical applications and guided homework assignments, you’ll be expected to create datasets, develop and train neural networks, explore your own media collections using existing state-of-the-art deep nets, synthesize new content from generative algorithms, and understand deep learning’s potential for creating entirely new aesthetics and new ways of interacting with large amounts of data. 

The online course is free ($10 a month for premium service) - you can find out more here

via devopsreactions.tumblr.com

by kevit

via cote.io

New OpenStack market-sizing and -forecast from old pals at 451:

• Al & Jay say $1.8bn in 2016, going to $5.4bn in 2020.
• Public cloud dominates now, but is expected to switch – “[public cloud providers are] 49% of total OpenStack revenue in 2015. However, we expect OpenStack private cloud service provider revenue to exceed public cloud providers by 2019.”

How they bucket-ize:

451 Research’s Market Monitor focuses on 56 vendors that provide direct OpenStack offerings, including products, services and turnkey offerings around OpenStack deployment and management, different distributions of OpenStack, service providers and training services. Although we do consider some vendors with integrated hardware, systems and software offerings based on OpenStack, our market-sizing estimate does not include hardware-centric revenue, nor does it include revenue from indirect third-party vendors, such as those in storage or software-defined networking.

Source: OpenStack-related business models to exceed $4bn by 2019

via zwischenzugs.wordpress.com

Overview

Docker is extremely popular with developers, having gone as a product from zero to pretty much everywhere in a few years.

I started tinkering with Docker three years ago, got it going in a relatively small corp (700 employees) in a relatively unregulated environment. This was great fun: we set up our own registry, installed Docker on our development servers, installed Jenkins plugins to use Docker containers in our CI pipeline, even wrote our own build tool to get over the limitations of Dockerfiles.

I now work for an organisation working in arguably the most heavily regulated industry, with over 100K employees. The IT security department itself is bigger than the entire company I used to work for.

There’s no shortage of companies offering solutions that claim to meet all the demands of an enterprise Docker platform, and I seem to spend most of my days being asked for opinions on them.

I want to outline the areas that may be important to an enterprise when considering developing a Docker infrastructure.

If I’ve missed anything or you have any comments get in touch below or tweet @ianmiell

Images

Registry

You will need a registry. There’s an open source one (Distribution), but there’s numerous offerings out there to choose from them if you want to pay for an enterprise one.

• Does this registry play nice with your authentication system?
• Does it have a means of promoting images?
• Does it have role-based access control?
• Does it cohere well with your other artifact stores?

Image Scanning

An important one.

When images are uploaded to your registry, you have a golden opportunity to check that they conform to standards. For example, could these questions be answered:

• Is there a shellshock version of bash on there?
• Is there an out of date ssl library?
• Is it based on a fundamentally insecure or unacceptable base image?

Static image analysers exist and you probably want to use one.

Image Building

How are images going to be built? Which build methods will be supported and/or are strategic for your organisation? How do these fit together?

Dockerfiles are the standard, but some users might want to use S2I, Docker + Chef/Puppet/Ansible or even hand-craft them.

• Which CM tool do you want to mandate (if any)
• Can you re-use your standard governance process for your configuration management of choice?
• Can anyone build an image?

Image Integrity

You need to know that the images running on your system haven’t been tampered with between building and running.

• Have you got a means of signing images with a secure key?
• Have you got a key store you can re-use?
• Can that key store integrate with the products you choose?

Third Party Images

Vendors will arrive with Docker images expecting there to be a process of adoption.

• Do you have a governance process already for ingesting vendor technology?
• Can it be re-used for Docker images?
• Do you need to mandate specific environments (eg DMZs) for these to run on?
• Will Docker be available in those environments?

SDLC

If you already have software development lifecycle (SDLC) processes, how does Docker fit in?

• How will patches be handled?
• How do you identify which images need updating?
• How do you update them?
• How do you tell teams to update?
• How do you force them to update if they don’t do so in a timely way?

Secrets

Somehow information like database passwords need to be passed into your containers. This can be done at build time (probably a bad idea), or at run time.

• How will secrets be managed within your containers?
• Is the use of this information audited/tracked and secure?

Base Image?

If you run Docker in an enterprise, you might want to mandate the use of a company-wide base image:

• What should go into this base image?
• What standard tooling should be everywhere?
• Who is responsible for it?

Security and Audit

The ‘root’ problem

By default, access to the docker command implies privileges over the whole machine. This is unlikely to be acceptable to most sec teams in production.

• Who (or what) is able to run the docker command?
• What control do you have over who runs it?
• What control do you have over what is run?

Solutions exist for this, but they are relatively new.

Monitoring what’s running

A regulated enterprise is likely to want to be able to determine what is running across its estate. What can not be accounted for?

• How do you tell what’s running?
• Can you match that content up to your registry/registries?
• Is what is running up to date?
• Have any containers changed critical files since startup?

Forensics

When things go wrong people will want to know what happened. In the ‘old’ world of physicals and VMs there were a lot of safeguards in place to assist post-incident investigation. A Docker world can become one without ‘black box recorders’.

• Can you tell who ran a container?
• Can you tell who built a container?
• Can you determine what a container did once it’s gone?
• Can you determine what a container might have done once it’s gone?

Operations

Logging

Application logging is likely to be a managed or controlled area of concern:

• Do the containers log what’s needed for operations?
• Do they follow standards for logging?
• Where do they log to?

Orchestration

Containers can quickly proliferate across your estate, and this is where orchestration comes in. Do you want to mandate one?

• Does your orchestrator of choice play nicely with other pieces of your Docker infrastructure?
• Do you want to bet on one orchestrator, hedge with a mainstream one, or just sit it out until you have to make a decision?

Operating System

Enterprise operating systems can lag behind the latest and greatest.

• Is your standard OS capable of supporting all the latest features? For example, some orchestrators and Docker itself require kernel versions or packages that may be more recent than is supported. This can come as a nasty surprise…
• Which version of Docker is available in your local package manager?

Development

Dev environments

• Developers love having admin. Are you ready to effectively give them admin with Docker?
• Are their clients going to be consistent with deployment? If they’re using docker-compose, they might resent switching to pods in production.

CI/CD

Jenkins is the most popular CI tool, but there’s other alternatives popular in the enterprise.

• What’s your policy around CI/CD plugins?
• Are you ready to switch on a load of new plugins PDQ?
• Does your process for CI cater for ephemeral Jenkins instances as well as persistent, supported ones?

Infrastructure

Shared Storage

Docker has in its core the use of volumes that are independent of the running containers, in which persistent data is stored.contains, in which

• Is shared storage easy to provision?
• Is shared storage support ready for increased demand?
• Is there a need for shared storage to be available across deployment locations?

Networking

Enterprises often have their own preferred Software Defined Networking solutions, such as Nuage, or new players like Calico.

• Do you have a prescribed SDN solution?
• How does that interact with your chosen solutions?
• Does SDN interaction create an overhead that will cause issues?

aPaaS

Having an aPaaS such as OpenShift or Tutum Cloud can resolve many of the above questions by centralising and making supportable the context in which Docker is run.

• Have you considered using an aPaaS?
• Which one answers the questions that need answering?

Cloud Providers

If you’re using a cloud provider such as Amazon or Google:

• How do you plan to deliver images and run containers on your cloud provider?
• Do you want to tie yourself into their Docker solutions, or make your usage cloud-agnostic?

Hey, what about x?

Get in touch: @ianmiell
My book Docker in Practice

Get 39% off with the code: 39miell

via blog.jupyter.org

Learning the lessons of the Jupyter Notebook

It’s been a long time in the making, but today we want to start engaging our community with an early (pre-alpha) release of the next generation of the Jupyter Notebook application, which we are calling JupyterLab.

At the SciPy 2016 conference, Brian Granger and Jason Grout presented (PDF of talk slides) the overall vision of the system and gave a demo of its current capabilities, which are rapidly evolving and improving:

Notebooks and terminals and kernels, oh my! #jupyterlab about to make my life way better at #scipy2016pic.twitter.com/BSxjRlMomu

— Chris Holdgraf (@choldgraf) July 14, 2016

JupyterLab captures a lot of what we have learned from the usage patterns of the Notebook application over the last 5 years and seeks to build a clean and robust foundation that will let us not only offer an improved user interface and experience, but also a flexible and extensible environment for interactive computing.

In reality, even today’s “Jupyter Notebook” is a bit of a misnomer: the Notebook application includes not only support for Notebooks but also a file manager, a text editor, a terminal emulator, a monitor for running Jupyter processes, an IPython cluster manager and a pager to display help. And that is just what ships “out of the box”, without counting the many third-party extensions for it. This rich toolset evolved organically, driven by the needs of our users and developers, even if we kept the increasingly ill-fitting “Notebook” name for the whole thing.

But the underlying code was showing its age: it wasn’t the cleanest to extend and many APIs that were somewhat experimental and incomplete had become “official” by virtue of being used in the wild. Providing a more responsive and flexible UI atop the current codebase was difficult. So, in a collaborative effort between the Jupyter team, Tech at Bloomberg and Continuum Analytics, we set out to build a next-generation architecture to support all of the above tools, but with a flexible and responsive UI, offering user-controlled layout that could tie together our tools under a single roof.

A detailed account of this collaboration between our teams is available at the Tech at Bloomberg Blog.

A glimpse of JupyterLab

While the system is still in alpha state and many features are missing, we can already see the kinds of user experiences it enables. Here we see how you can arrange a notebook next to a graphical console (a web-based version of our standalone QtConsole) atop a terminal that is monitoring the system, while keeping the file manager on the left:

With a few clicks, you can reorganize the workspace to expose the Command Palette to access all JupyterLab functions, keeping the graphical console side-by-side with a text editor containing a Python script that you then %run from the console, producing inline figures:

These examples illustrate how the new system, based on Continuum’s flexible PhosphorJS framework, gives us the foundation for a richer, cleaner UI. JupyterLab adapts easily to multiple workflow needs, letting you move from a Notebook/narrative focus to a script/console one. It exposes the Jupyter tools we all use daily and will let both the core team and the entire community develop many new ones that take advantage of the Jupyter architecture. The entire JupyterLab is built as a collection of plugins that talk to kernels for code execution and that can communicate with one another. We hope the community will develop many more plugins for new use cases that go far beyond the basic system.

Even in its current alpha state we are very excited about the possibilities. If you are willing to play with very early code, you can follow the instructions on the repo and join us in helping test and refine the system.

This effort is the fruit of an open collaboration between our industry partners at Bloomberg and Continuum and the Jupyter Team anchored at UC Berkeley/LBNL and CalPoly, funded by the Helmsley Trust, the Gordon and Betty Moore Foundation and the Alfred P. Sloan Foundation. We are extremely grateful for this support, and we hope to see in the future many more examples of similar partnerships between academia, philantrophic funders and industry.

via vanishingnewyork.blogspot.com

via prostheticknowledge.tumblr.com

Fusion4D: Real-time Performance Capture of Challenging Scenes

Latest from Microsoft Research reveals state of the art developments in videogrammetry, the method of capturing realtime virtual 3D form - in this case, how their technology can even capture intricate and difficult details:

We contribute a new pipeline for live multi-view performance capture, generating temporally coherent high-quality reconstructions in real-time. Our algorithm supports both incremental reconstruction, improving the surface estimation over time, as well as parameterizing the nonrigid scene motion. Our approach is highly robust to both large frame-to-frame motion and topology changes, allowing us to reconstruct extremely challenging scenes. We demonstrate advantages over related real-time techniques that either deform an online generated template or continually fuse depth data nonrigidly into a single reference model. Finally, we show geometric reconstruction results on par with offline methods which require orders of magnitude more processing time and many more RGBD cameras. 

Link