/* ---- Google Analytics Code Below */

Saturday, August 19, 2017

More Replenishment

More looks at the retail replenishment angle,  with discussion,  also see Amazon's use of this idea:

Will the ‘SmartStockUp’ program drive replenishment sales for Boxed?   by Matthew Stern

Consumers buy staple products in bulk to avoid running out, but it still invariably happens. E-tailer Boxed, whose business model is often said to resemble an e-commerce take on wholesale clubs like Costco, is trying a new method to remind customers to replenish.

Boxed has piloted a new program called SmartStockUp with about a quarter of its user base and is now expanding it to all its B2C customers, according to TechCrunch. SmartStockUp uses information it has about customer habits, such as purchase data, along with broader restocking trends to anticipate when customers will need to replenish a particular item. Boxed then displays a “Need this Now” or “Need this Soon” recommendation when the customer visits the website. .... " 

Introduction to GAN Engine, Applications

Good piece I have sent forward as an introduction to Generative Adversarial Networks (GANs)  Especially with regard to useful applications.

Posted by Luba Belokon in DSC:

GAN Engine and Applications: 
Generative adversarial networks (GANs) are a class of neural networks that are used in unsupervised machine learning. They help to solve such tasks as image generation from descriptions, getting high resolution images from low resolution ones, predicting which drug could treat a certain disease, retrieving images that contain a given pattern, etc.    Our team asked a data scientist, Anton Karazeev, to make the introduction to GANs engine and their applications in everyday life.  ....  " 

More on the Jinan Project


China Set to Launch the World’s First Quantum Communication Network  in Futurism. 

China has successfully tested the Jinan Project, and is set to begin using the system by the end of August. This marks a world milestone in the development of quantum technology, and identifies China as one of the world leaders in the field.  .... "

China Advances in Quantum Computiing

Quantum computing followed here for years, before China was even mentioned in the space:

China has successfully tested the Jinan Project, and is set to begin using the system by the end of August. This marks a world milestone in the development of quantum technology, and identifies China as one of the world leaders in the field.

By the end of August, China plans to rollout the Jinan Project — the world’s first unhackable computer network, which is based on quantum principles. The project uses the city of Jinan as a quantum computer hub that boosts the Beijing-Shanghai quantum network due to its central geographical position between the two larger cities.  ... " 

Apple and Augmented Reality

We have seen AR demonstrated on the phone for a long time, but it has not broken through to real value or common use.   Is that about to change?

Apple Bets the Future of Augmented Reality will be on your Phone
THE AR REVOLUTION will come not with a bang, but a tape measure. At least, that seems to be the lesson so far from ARKit, Apple’s new augmented-reality platform.

ARKit lets developers build AR apps—which integrate digital experiences into the physical world via iPhone or iPad, a la Pokemon Go. Those apps will be available to consumers when iOS 11 arrives in September. But developers have started tinkering—creating tools that let you see how furniture fits in a room, or quickly calculate the area of your kitchen. Compared to the likes of Magic Leap or Google Glass, these apps are simple, almost trivial. But that smallness might be precisely what makes them so potentially huge.

Matthew Miesnieks, a VC who led AR research and development efforts at Samsung, calls ARKit “the biggest thing that’s happened to the AR industry since it began,” and he’s not alone in his enthusiasm. By getting AR in the hands of millions of iPhone users, Apple is poised to become the world’s most powerful and popular purveyor of augmented-reality apps. And by opening up its developers’ kit, it’s powering hundreds of experiments into what, precisely, this medium is good for. ....  "

Friday, August 18, 2017

WorkFusion

Continue to look at means to automate the specific process of business. Making it better, faster, cheaper. And ultimately include intelligent and cognitive elements to streamline and have these systems adapt to context.  I see that Deloitte is a partner.

Brought to my attention: WorkFusion.

NEW YORK, July 18, 2017 — WorkFusion today released Intelligent Automation 2017 Sunbird, a comprehensive product update that enables WorkFusion customers to automate more work with less effort and easily scale the benefits of automation across a business. Sunbird brings more of WorkFusion’s vision of low-code RPA and cognitive automation into reality and makes its flagship product, Smart Process Automation (SPA), faster, simpler, more scalable and more secure.

“We created Intelligent Automation 2017 Sunbird update in collaboration with our customers,” says WorkFusion’s Head of Product, Mikhail Abramchik. “Their feedback inspired us to build features and improvements to make our product easier to use and more efficient.”

Key customer-driven benefits of the Sunbird update release include:
• Cognitive bots that learn 8 times faster, delivering automation quicker
• More out-of-the-box use cases for faster, easier deployment
• RPA recorder to create bots without coding
• RPA API and universal driver means 50% less effort for surface and Excel automation
• Best-in-class password vault for enterprise-grade security
• Load balancing to eliminate redundancy, which reduces infrastructure cost

These improvements translate to a better experience for WorkFusion’s customers across the boardfrom business users to developers to the C-suite. A simpler process for setting up cognitive bots allows business users to automate more with less coding, while enhanced developer tools and OCR plugin reduce the time and resources developers spend on RPA coding. From an executive perspective, all these enhancements help to scale Intelligent Automation across an entire operation to meet pressing transformation goals, which means increased ROI and better service delivery.

Says Abramchik: “We’re proud to say that Intelligent Automation 2017 Sunbird offers more possibilities to transform and reinvent businesses than any other RPA or cognitive automation product on the market today.”

About WorkFusion
WorkFusion Intelligent Automation empowers enterprise operations to digitize. WorkFusion delivers the quick wins of robotic process automation (RPA) and the longer-term transformation and ROI of AI-powered cognitive automation through products purpose-built for enterprise operations. The world’s leading global banks, insurance, healthcare companies, and BPO firms use WorkFusion Intelligent Automation to become leaner, more productive and agile.  .... " 

Disruption is not the Measure

Sometimes all I hear is disruption, but it is not all there is, and itself is not a measure.  We made money for many years on improvement alone.  So look for disruption, but include hints to value along the way.  By Deloitte in the WSJ:

Disruption Is Not the Key to Winning
Companies can create competitive advantage by leveraging digital technologies to provide exceptional experiences for customers. Six enablers can help.

“Disruptor” is a term used frequently to describe successful modern companies. Airbnb is often credited with disrupting the lodging industry, Uber is cited as a disruptor of the transportation business, and Amazon is widely seen as disrupting retail. Yet while these companies have certainly transformed their industries through innovative business models, disruption isn’t the yardstick for measuring success.  ... " 

Stats on Privately Funded Drone Companies

CBInsights: A look at privately funded drone companies.   " ... .While the US accounts for nearly two-thirds of all drone investment activity, there are at least 11 other countries making bets in the space  ....  Since 2012, private drone companies have raised over $1.6B across 300+ equity investments. Drones make up the second largest sub-category in IIoT (Industrial Internet of Things) and account for roughly 20% of total deal activity.

Three of the earliest private drone companies and largest in terms of total funding are 3D Robotics, Airware, and RedZone Robotics. While they lead in the US, expanding internationally has been a challenge because of the complex international regulatory environment, both at country and local levels.  ... " 

Mixed Reality Partner Program

Received:

Introducing the Mixed Reality Partner Program

If your company is a digital agency, systems integrator, or solution provider, the Mixed Reality Partner Program can equip your company with the skills to successfully drive commercially-focused mixed reality solutions for your clients. .... " 

Georgia Institute of Technology Robotarium

Magnus Egerstedt, executive director for the Institute for Robotics and Intelligent Machines at the Georgia Institute of Technology, in the new Robotarium. Been a while since I have been to Ga Tech's lab spaces, we visited while I ran tech in our company's innovation centers.  Impressive then, and they continue to improve.   Had seen the remote lab idea suggested before, but did not know they were doing this.   Need to get there.

This Robot Lab Has No Idea What Its Robots Are Doing    By The Wall Street Journal 
Magnus Egerstedt, executive director for the Institute for Robotics and Intelligent Machines at the Georgia Institute of Technology, in the new Robotarium. 

In the Robotarium, a 10-square-foot table inside the Atlanta laboratory of Magnus Egerstedt, executive director of the Georgia Institute of Technology’s Institute for Robotics and Intelligent Machines, up to 50 ground bots and 20 aerial copters can be remotely controlled by researchers often from other schools, or even foreign countries ....

The Georgia Institute of Technology's (Georgia Tech) Institute for Robotics and Intelligent Machines supports the Robotarium, an arena where scientists can run experiments on remote-controlled machines.

Georgia Tech's Magnus Egerstedt says unpredictability is a regular feature of the Robotarium, where swarms of ground and airborne robots are put through their paces. Researchers use these swarms to test search-and-rescue scenarios, model flight formations for the U.S. Department of Defense, and predict the interactions of fleets of autonomous cars. Cameras located throughout the arena record the trials so the researchers conducting them can see the results of their experiments.  ... " 

Fingerprint Analysis Scoring

Recall seeing this problem as a classic pattern recognition issue,  closer to being solved.  Note this addresses knowledge content in a pattern, where else might this be used? Anywhere you might want to prioritize an action.

Scientists Automate Key Step in Forensic Fingerprint Analysis
By U.S. National Institute of Standards and Technology 

Researchers at Michigan State University and the U.S. National Institute of Standards and Technology have developed an algorithm that automates the fingerprint-analysis process.

The researchers trained the machine-learning algorithm on data from 31 fingerprint experts who had analyzed 100 latent prints each, scoring the quality of each on a 1-to-5 scale. The prints and their scores were used to train the algorithm to determine how much information a latent print contains.

The researchers tested the algorithm by having it score a group of new latent prints. The team submitted those scored prints to Automated Fingerprint Identification System software connected to a database of more than 250,000 rolled prints. The researchers found the scoring algorithm performed slightly better than the average of human examiners involved in the study.

They next want to test the system on a larger dataset, which will enable them to improve the algorithm's performance and more accurately measure its error rate.

From U.S. National Institute of Standards and Technology

(More details at the link)

Every Possible Song Has Some Non-Zero Streaming Value

When looking at the streaming concoctions of smart speakers. I quickly noticed the increase of so called 'mood' music playlists, aimed at providing backgrounds for sleeping, studying, working, concentrating, cooking, vanity, etc.   Unclear if any of this works.  But the user, including myself, is quick to bite and try.  And a slightly manipulated search can quickly get you there.   I was not the first to notice this.  Even giants like Spotify can create such 'fake' music.

One company is already commissioning  a version of 'Happy birthday to your name', for every possible given name.  Starting with the most common.  The one for 'Matthew' has already been included in 400,000 playlists.  What is 'fake' is debatable.  Perhaps written by machine and performed by studio artists.   Or created out of bits and snippets of real music.  Or created in its entirety by AI.   Regardless, cheaper to license than music by real artists, and much less expensive than from real, famous composers and artists.   Will we ultimately have to specify detect and specify 'real' music?

The accusation against Spotify in particular is written about in the NYT.  And an article about related efforts in Vulture.

Thursday, August 17, 2017

1,600 Alexas put in AZ State Engineering Dorm Rooms

This could be a very big deal, if loaded with the right software.  And for prompting potential forward looking skill development.   Think smarter.  Will continue to follow what come out of that.  Anyone from the AZ State project, please do contact me.

Amazon will put 1,600 Echo Dots in Arizona State University’s engineering student dorms ... Talk about real-world application   By Dani Deahl on August 17, 2017

This year, students at Arizona State University will have the opportunity to live in college housing with Echo Dots in a program that encourages engineering students to practice voice user interface development skills on consumer hardware.

ASU has built a new work / live space for first-year engineering students called the Tooker House, and those moving into its residence hall will be able to opt in to the program and receive an Echo Dot for their dorm room. The school says Amazon has donated 1,600 Echo Dots and is also providing developer kits to help add the technology to ASU’s existing engineering curriculum.

Outside of Tooker House, any student in ASU’s engineering school can enroll in one of three upcoming fall courses that teach concepts like voice user interface development, which includes Alexa skills. The students will be encouraged to independently build Alexa skills outside the classroom using the Alexa Skills Kit (ASK), which can ideally be incorporated into student project programs, or solve needs in the local community. ... " 

Respecting Absurd Ideas

Always liked the idea of exploring the edge of possibilities.  Here a look at how to exploit absurd suggestions.   Which break down into Respect the idea, and Play with it.   I would add,  always use the promoter of the idea as the 'expert' in its application .... they can school you in its application, or even why it won't work in context.   Stretch it.  Explore with analogies.   Absurdity can be good.

strategy+business: Corporate Strategies and News Articles on Global Business, Management, Competition and Marketing
 Two Simple Concepts for Getting the Most from Absurd Ideas ... " 

Classifications

We all like to classify things,  its efficient.   Mathematicians just like to classify more formally.  And we use that in data science so we can make use of the results.  It is learning.  An article in Quanta Mag talks this:  Why Mathematicians Like to Classify Things.

Agile in the C-Suite

In Bain & Company:

How to Make Agile Work for the C-Suite   By Eric Garton, Andy Noble

Many companies are attempting a radical — and often rapid — shift from hierarchical structures to more agile environments, in order to operate at the speed required by today’s competitive marketplace. Companies like ANZ, the Australian-based banking giant, have made explicit commitments to adopt agile principles, while others like Zappos, are on the bleeding edge of organizational transformation. Many stopping points exist along the continuum from hierarchy to holacracy. To successfully transform to a more agile enterprise, companies must make conscious choices about where and how to become agile. They have to decide where to adopt agile principles and mindsets, where to use agile problem-solving methodologies to dynamically address strategic and organizational challenges, and where to more formally deploy the full agile model, including self-managed teams.

At Bain & Company, we do not believe that companies should try to use agile methods everywhere. In many functional areas, such as plant maintenance, purchasing, sales calls, or accounting, more traditional structures and processes likely will deliver lower cost, more repeatable outcomes and more scalable organizations. Sorting through every function and every part of your company’s operating model to determine which parts of the agile playbook to adopt requires some deep thinking. It also means you have to figure out how to make the agile and traditional parts of your organization effectively operate with one another. This takes time. .... " 

Inc Looks at Skillz

I see that former colleague Andrew Paradise's game company continues to grow.  An article in Inc has lots of detail:

How the No. 1 Fastest-Growing Company in America Turned Video Games Into a $54 Million Cash Cow    How did America's fastest-growing company learn the skills to conquer? Countless hours of e-sports..... "    By David Whitford  Inc Editor-at-large@davidwhitford

Wednesday, August 16, 2017

Facebook Messenger's M

Been asked to go back and look at some of the AI/Assistant components of Facebook M, just launched in the UK after some time in the US.  Facebook's work in the AI and machine learning space continues to expand.

In particular look at their work on dialog research,  covered in the tags below. Ultimately handling dialog intelligently will be key in assistance.  The operationally multilingual approach is also rare.

Everything You Need to Know About Facebook Messenger's M Assistant
Facebook's M assistant can now suggest video or voice calls   By Christian de Looper in DigitalTrends. 

Notice a little "M" logo near the textbox in Facebook Messenger? It's the social media network's artificially intelligent assistant that's meant to offer suggestions in your conversations -- similar to how Google Assistant can jump in to help in conversations via Allo. With the assistant officially rolled out across the U.S., it is now available in both English and Spanish to all Messenger users in the country.  

M is now able to provide suggestions in Spanish -- like sending payments or sharing your location -- and the feature is gradually rolling out in Mexico for iOS and Android, Facebook told Digital Trends. If you'd like to receive M Suggestions in Spanish, you can change the preferred language to Spanish.

M itself is powered by Facebook-built AI, and so far works by popping up in your chats when it detects that it can help with something. For example, if you mention going somewhere, M can detect this and will offer up a quick way to book a ride with Uber or Lyft (you can set your preference) -- all without leaving the Messenger app.

Before diving into what Facebook's M can do, there are a few things to keep in mind. Most importantly, Facebook is still working on M and it's likely the assistant will evolve a lot over the next few months and years. It will also roll out globally at some point as Facebook says users across the globe will "start seeing a redesigned way to compose messages." If you want nothing to do with this AI revolution, or if you think M makes your conversations too cluttered, you can mute it by heading to Messenger's settings -- tap on the person icon on the top-right of Messenger's main page, then tap M settings and toggle Suggestions off.   ... ." 

My coverage of this and many other personal assistants.

Making Calls with Google Home

I just complained about this, and to be clear it has still not rolled out to me.  But apparently is on its way.    Some of the apparent differences are described.    Have been using the Amazon Echo for this since the start of this year,  most often using it as an in-house intercom, which it does well. This for me most useful aspect of Echo calling is not mentioned in the article below.  GHome does not permit specific device intercom or two way service,  just outgoing  phone calls.     The Verge talks the specifics of trying this, will be doing that:

How to make phone calls with Google Home
It’s more flexible than making calls with Amazon’s Alexa  by Chris Welch 

Google’s smart speaker can now pull double duty as a phone for voice calls. The company just confirmed that it’s rolling out Google Home’s calling feature in the US and Canada beginning today. Users can dial anyone in their contacts and local businesses for free — so long as the call recipient is in one of those two countries. The calling feature was first announced back in May.

In turning its speaker into a phone, Google is taking another step to challenge Amazon and its Echo devices, which introduced calling and messaging features earlier this year. But the two companies take a significantly different approach in how the feature actually works and who you’re able to communicate with.   ... " 

See also: Google Home calling support page.

Amazon Pays for Popular Skills

One time small payment, it seems,  so does not seem sustainable.  The incentive is to get you to go back to the device.  Classic engagement.   While I have been impressed by the number of skills available, their quality and interest still seems low.  A stronger memory of your visits ... personalization ... gamification...  Competition ....   real surprise in results.     Would all seem to be better suited to generate an ongoing repeat experience.

Amazon will pay more developers who make popular Alexa skills
Payment is being expanded into more categories
by Thuy Ong@ThuyOng   In the Verge

Gartner Ranks AI as top Tech to Watch

No surprise, Gartner ranks AI as the top technology to watch in coming years.  Includes their time-hype graph of many other technologies.  Via SiliconAngle.      By Mike Wheatley

Machine Teaching

Out of Microsoft Research.  O'Reilly Summarises.

Machine Teaching: A New Paradigm for Building Machine Learning Systems
Patrice Y. Simard, Saleema Amershi, David M. Chickering, Alicia Edelman Pelton, Soroush Ghorashi, Christopher Meek, Gonzalo Ramos, Jina Suh, Johan Verwey, Mo Wang, John Wernsing

The current processes for building machine learning systems require practitioners with deep knowledge of machine learning. This significantly limits the number of machine learning systems that can be created and has led to a mismatch between the demand for machine learning systems and the ability for organizations to build them. We believe that in order to meet this growing demand for machine learning systems we must significantly increase the number of individuals that can teach machines. We postulate that we can achieve this goal by making the process of teaching machines easy, fast and above all, universally accessible.

While machine learning focuses on creating new algorithms and improving the accuracy of "learners", the machine teaching discipline focuses on the efficacy of the "teachers". Machine teaching as a discipline is a paradigm shift that follows and extends principles of software engineering and programming languages. We put a strong emphasis on the teacher and the teacher's interaction with data, as well as crucial components such as techniques and design principles of interaction and visualization.

In this paper, we present our position regarding the discipline of machine teaching and articulate fundamental machine teaching principles. We also describe how, by decoupling knowledge about machine learning algorithms from the process of teaching, we can accelerate innovation and empower millions of new uses for machine learning models.

QR Codes in China

Have not been using or reading about the use of QR codes of late, interesting how they are being used in China. See also links to previous pieces on QR.    Some good, straight forward examples. Key too is how ready the customer base is willing to scan them.

16 Ways QR Codes are Being Used in China
by Connie Chan,   In Andreessen Horowitz Blog

We’ve talked a lot about the rise of QR codes in Asia, but they may now finally be moving from being a “joke” to being more widely adopted in other places as well. Simply put, QR codes let you hyperlink and bookmark the physical world. Just as UPC barcodes allow machine-readable scanning of data (e.g., price) on items in stores, QR codes are a barcode-like vector between online and offline information. And unlike NFC (near-field communication), which is used for reading smart cards, keycards, and contactless payments, QR codes can be easily accessed by any phone in the world that has a camera. They enable everything from online to offline (O2O) marketplaces, which are huge in China, to augmented reality.

Some of the more obvious use cases for them include things like adding a WeChat friend in real life (IRL); subscribing to a WeChat official account (often representing media, stores, people, and others); paying a street vendor or at a convenience store; connecting to wi-fi in a shop; getting additional content from a magazine article; and learning more about styling or the brand from a clothing label. But there are also a number of less-obvious (or not as well covered) uses in China, which I share below, because they show the range of what’s possible everywhere when QR codes disintermediate existing use cases… and enable new ones.  .... 

Tuesday, August 15, 2017

Assistants Using Wikipedia

A conversation at Wikipediocracy discusses how personal assistants like Alexa, Google Home and Siri use and attribute (or not)  knowledge from the Wikipedia.   Back to the complicated world of licenses and copyrights.   Was also re-introduced to the considerable undercurrent of WP authors complaining about how their work is used.

Autonomous Video for the Home

IEEE Spectrum on 'friendly'  home robotics.   Still awaiting some more detailed looks at its in home use. Continue to cover this, but given the cost will likely not dive in without some convincing.

Kuri Robot Brings Autonomous Video to a Home Near You
Mayfield Robotics improves its home robot, Kuri, adding track wheels, structural updates, and “Kuri Vision,” an autonomous home video program

Most home robots are designed primarily for convenience and function. Not Kuri. Silicon Valley startup Mayfield Robotics designed Kuri specifically to be an adorable home companion. And that means it needed to have one quality you won’t find in most robotic vacuums and other home bots: cuteness. 

Mayfield introduced Kuri earlier this year at the Consumer Electronics Show in Las Vegas. Since then, the Mayfield team has made several updates to the robot. The most significant one is the home video feature called “Kuri Vision,” which allows Kuri to take video autonomously.

To do that, Kuri has two high definition 1080p cameras, one behind each eye. These cameras take videos intermittently throughout the day, capturing candid moments. You can then review those clips through the app, which runs on iOS and Android, and choose which ones you like best. Then Kuri’s machine learning and image processing kicks in: Based on which images you favorite or delete, Kuri learns to take videos that you’ll like. .... " 

The Dataset that Transformed AI Research

As in the emergence of Big Data, it has been pointed out that the large amount of the right kind of data can make the difference.  And it has, opening an entire industry of image recognition and understanding, by sculpting algorithms to interpret them.  Re opening the idea of AI, that had slumbered since the 80s.

It's not about the Algorithm  

The data that transformed AI research—and possibly the world  In QZ by Dave Gershgorn   @davegershgorn

In 2006, Fei-Fei Li started ruminating on an idea.

Li, a newly-minted computer science professor at University of Illinois Urbana-Champaign, saw her colleagues across academia and the AI industry hammering away at the same concept: a better algorithm would make better decisions, regardless of the data.

But she realized a limitation to this approach—the best algorithm wouldn’t work well if the data it learned from didn’t reflect the real world.

Her solution: build a better dataset.

“We decided we wanted to do something that was completely historically unprecedented,” Li said, referring to a small team who would initially work with her. “We’re going to map out the entire world of objects.”

The resulting dataset was called ImageNet. Originally published in 2009 as a research poster stuck in the corner of a Miami Beach conference center, the dataset quickly evolved into an annual competition to see which algorithms could identify objects in the dataset’s images with the lowest error rate. Many see it as the catalyst for the AI boom the world is experiencing today. ... " 

Google View of AI: Intelligence and Scale

Via O'Reilly: 

Jeff Dean is a Google senior fellow in the Research Group, where he leads the Google Brain project. Here is a video (and slides) of a talk he gave, "Intelligent Systems with Large Scale Deep Learning." It's a decent intro to AI, with some hints about how Google expects AI to move forward. ....

Monday, August 14, 2017

Identifying Plant Species

Another example of using many images to train via AI.   Also describes the data needs for such a process, done via Deep Learning neural methods.

From the CACM: 
Digitizing plant specimens is opening up a whole new world for researchers looking to mine collections from around the world.

Computer algorithms trained on the images of thousands of preserved plants have learned to automatically identify species that have been pressed, dried and mounted on herbarium sheets, researchers report. ....  " 
Artificial Intelligence Identifies Plant Species for Science  In Nature 
" .... Bonnet's team had already made progress automating plant identification through the Pl@ntNet project. It has accumulated millions of images of fresh plants — typically taken in the field by people using its smartphone app to identify specimens.

Researchers trained similar algorithms on more than 260,000 scans of herbarium sheets, encompassing more than 1,000 species. The computer program eventually identified species with nearly 80% accuracy: the correct answer was within the algorithms’ top 5 picks 90% of the time. That, says Wilf, probably out-performs a human taxonomist by quite a bit. .... " 

Infinite Pool Tables

We actually used this idea for solving cleaning coverage problems.   A rare case where advanced topology math principles came into play in industry.  This was the kind of math that I always liked, not too abstractly symbolic, but visually interesting.

New Shapes Solve the Infinite Pool-Table Problem

NASA and Virtual and Augmented Realities

NASA has been known for experimenting in this space, I have played with a few of their published capabilities.  Here is an overview of their future plans.

NASA’S Next Spacecraft  may launch from virtual and augmented realities.     By Dyllan Furness in Digital Trends.  ..... "

Time Series Insights in Azure

Time Series Insights PREVIEW
Instantly explore and analyze time-series data in IoT solutions
Azure Time Series Insights is a fully managed analytics, storage, and visualization service that makes it simple to explore and analyze billions of IoT events simultaneously. It gives you a global view of your data, letting you quickly validate your IoT solution and avoid costly downtime to mission-critical devices by helping you discover hidden trends, spot anomalies, and conduct root-cause analyses in near real-time.  .... " 

(Start for free at link, taking a look at the root cause example, always a great analytics place to start, because it concerns everyone) 

Cortana Predicting Future Travel Plans

Been watching virtual assistants for some time.    The key appears to be getting as many as possible out there,  on many kinds of hardware,  keep them cheap and for attentive use with voice,  something you use every day,  but not as part of your phone and computer,  link it to a few key common entertainment functions (like music), open them to external skill development and keep coming out with new capabilities.  

Cortana has been available for some time now, but I remain unimpressed.  Note I am a satisfied Windows 10 user, always up to date, so it should be an easy sell. It should be linking to all my office functions, building intelligence into their use, but does not.  I spend more time shutting Cortana off than using it.  Still awaiting its implementation on a stand-alone on what looks to be a premium price  a Harman-Kardon device.

Now word is out that Cortana will predict your future travel plans.    Nice idea, but still not something I do every day,  will try it, but suggest you would be better to get it out in many users and developers hands.   Soon.

More of my coverage of personal assistants. 

Video Detecting Infection Patterns

Looking for patterns in healthcare using video that lead to infection.

Researchers use AI to monitor hospital staff hygiene
The technology could be used to reduce rates of hospital-acquired infections.

By Mallory Locklear,   @mallorylocklear  in Engadget. 

Hospital-acquired infections are a pesky problem and around one in 25 hospital patients have at least one healthcare-associated illness at any given time. To combat this issue, a research team based at Stanford University turned to depth cameras and computer vision to observe activity on hospital wards -- a system that could be used to track hygienic practices of hospital staff and visitors in order to spot behaviors that might contribute to the spread of infection. The work is being presented at the Machine Learning in Healthcare Conference later this week.  ... " 

Text Summarization

A good AI challenge that has useful applications.

An Algorithm Summarizes Lengthy Text Surprisingly Well
Training software to accurately sum up information in documents could have great impact in many fields, such as medicine, law, and scientific research.

by Will Knight  May 12, 2017  in Technology Review.

Who has time to read every article they see shared on Twitter or Facebook, or every document that’s relevant to their job? As information overload grows ever worse, computers may become our only hope for handling a growing deluge of documents. And it may become routine to rely on a machine to analyze and paraphrase articles, research papers, and other text for you.

An algorithm developed by researchers at Salesforce shows how computers may eventually take on the job of summarizing documents. It uses several machine-learning tricks to produce surprisingly coherent and accurate snippets of text from longer pieces. And while it isn’t yet as good as a person, it hints at how condensing text could eventually become automated .... " 

Sunday, August 13, 2017

Exoskeleton Controlled by Voice

Adding the voice to being reactive to movements is an interesting approach, another case of multi channel interaction and control.

This exoskeleton can be controlled using Amazon’s Alexa
‘Alexa, let’s go for a walk’     by James Vincent   @jjvincent

Amazon’s Alexa is available on a lot of devices, from lamps to alarm clocks to fridges. But robotics company Bionik Laboratories says it’s the first to add the digital assistant to a powered exoskeleton. The company has integrated Alexa with its lower-body Arke exoskeleton, allowing users to give voice commands like “Alexa, I’m ready to stand” or “Alexa, take a step.”

Movement of the Arke, which is currently in clinical development, is usually controlled by an app on a tablet or by reacting automatically to users’ movements. Sensors in the exoskeleton detect when the wearer shifts their weight, activating the motors in the backpack that help the individual move. For Bionik, adding Alexa can help individuals going through rehabilitation get familiar with these actions.   ... "

On a Bias about First Impressions

A first impression knowledge Bias?

Why our brains lead us astray when we take things at face value   Article by Diana Gitig 
A new book looks at how we overestimate what we can tell from a first impression.

" ... Professor Alexander Todorov’s new book, Face Value: The Irresistible Influence of First Impressions, is about much more than 19th-century pseudoscience. It’s about first impressions more generally. We all form them instantly—within 30-40 milliseconds, before we can consciously register even seeing a face. And we start exceptionally early on, probably at around seven months of age. We also seem to agree on these impressions, which makes the physiognomists’ promise so appealing.  .... " 

Addressing the Analogy Gap

Reminiscent of using humans as a peripheral, here determining high level relationships, then having the deep learning sort out the lower level patterns.  Points to Mechanical Turk, which we used this way.  At what point are the results general?  The comment about scale is key.  Also the mapping involved, very useful for generalizing and testing.

Crowdsourcing may have just helped close the "analogy gap" for computers    It's vexed computer scientists for decades, but a huge roadblock for true AI is falling    By Greg Nichols for Robotics

Researchers at Carnegie Mellon University (CMU) and the Hebrew University of Jerusalem in Israel have used crowdsourcing to teach computers to generate analogies so they can mine datasets to address new challenges by repurposing old concepts. "After decades of attempts, this is the first time that anyone has gained traction computationally on the analogy problem at scale," says CMU professor Aniket Kittur. The researchers hired participants via Amazon Mechanical Turk, tasking them to look through products on an innovation website and find analogous products from the same source. The participants noted which words caused them to link disparate products, mapping each pathway. Computers with deep-learning algorithms used these insights to analyze additional product descriptions and find new analogies. The researchers say this strategy can be used to customize computer programs to identify analogies between patent applications and literature on global problems.  ... " 

China and AI

In the NYT, word on China investments in AI: 

" ... China has laid out a development plan to become the world leader in artificial intelligence (AI) by 2030, with the goal of surpassing its rivals technologically and establishing a domestic industry worth nearly $150 billion. The policy, released by the State Council, is a statement of intent from the upper levels of China's government that the country will be investing heavily to ensure its companies, government, and military jump to the forefront of AI technology. The plan comes while China prepares a multibillion-dollar national investment initiative to support "moonshot" projects, startups, and academic research in AI.  ... "

Saturday, August 12, 2017

IFTTT and Honeywell

Continue to be impressed about how companies are working with IFTTT data streams, most recently brought to my attention:  Honeywell.  Plus other methods using the IFTTT Open Platform.  I have several examples in operation.

From our partners: 
“IFTTT’s network of partners is so extensive that that one connection allows us to get a lot of connections to third-party devices… IFTTT gave us a way to get there very quickly and very inexpensively.”          Scott Harkins   VP, Honeywell Connected Home

Video

Multivariate Regression

In DSC, a good piece explaining multivariate regression.  Not very technical, addresses both the data and the results.  Every manager should understand this simple approach.   A very common kind of problem you run into.   A good thing to walk through with decision makers and their real data.   Also leads you naturally to the next question, how do I determine which variables create the best predictive model?   The next step.

Bikes Existing Among Cars

Troublesome thought that bicycles might need to become so complex to exist in an ecosystem with automobiles.

Bikes May Have to Talk to Self-Driving Cars for Safety's Sake     by Margaret J. Krauss

Researchers envision bicycles communicating with autonomous vehicles so the latter can predict cyclists' movements. Waymo's self-driving autos have honed their predictive abilities over many simulated and actual driven miles, notes Waymo's Nathaniel Fairfield. Waymo's vehicles are programmed to pass bikes in compliance with state laws, or to wait if such action is impossible. Carnegie Mellon University professor Anthony Rowe wants bikes to feed data to cars. "We're trying to...put as much instrumentation on a bike as we can to see if we can predict how it's going to move in the future, so that it could, for example, signal a collision-warning system on a car," Rowe says. His team wants to collect as much information as possible to determine the precise and constant position of a bike in the world, and then determine the least amount of data a car requires from a cyclist for it to trigger an automatic braking system .... " 

Handbook of Neural Computation

Handbook of Neural Computation - 1st Edition - ISBN: 9780128113189, 9780128113196Below book looks good, but very pricey.    Consider a free version via MIT, also being continually updated, don't have a copy to compare its coverage.    In general today, a book on 'Deep Learning' would cover similar topics than one on 'Neural Computing',  but that is not necessarily implied by the titles.   Artificial neurons do learn by computing,  but the neural computing by itself does not mean learning, yet likely does today.

Handbook of Neural Computation 1st Edition,   Elsevier,2017
Posted by Sanjiban Sekhar Roy 

Handbook of Neural Computation explores neural computation applications, ranging from conventional fields of mechanical and civil engineering, to electronics, electrical engineering and computer science. This book covers the numerous applications of artificial and deep neural networks and their uses in learning machines, including image and speech recognition, natural language processing and risk analysis. Edited by renowned authorities in this field, this work is comprised of articles from reputable industry and academic scholars and experts from around the world. .... " 

Debating Statistical Significance

A considerable look, both technical and non-technical about statistical significance. Have it has been used, and how that is being re-considered.    The original title says this is a nerdy debate, I disagree, it is very important.  Having replicable significance is essential.

The case for, and against, redefining “statistical significance.” 
Updated by Brian Resnick

 There’s a huge debate going on in social science right now. The question is simple, and strikes near the heart of all research: What counts as solid evidence?

The answer matters because many disciplines are currently in the midst of a “replication crisis” where even textbook studies aren’t holding up against rigorous retesting. The list includes: ego depletion, the idea that willpower is a finite resource; the facial feedback hypothesis, which suggested if we activate muscles used in smiling, we become happier; and many more.

Scientists are now figuring out how to right the ship, to ensure scientific studies published today won’t be laughed at in a few years.

One of the thorniest issues with this question is statistical significance. It’s one of the most influential metrics to determine whether a result is published in a scientific journal.  .... " 

Telepresence Robots

These robots should be called 'minimal presence portable robotics',  we looked at them to test the idea of having someone who was very remote introduced to a team in a unique way.   They have lots of problems.  We discovered most of these in our tests.  The article is quite good and complete, covering several devices, and exposing many of their issues.

In particular that the remote person has to be trained on the device and wastes most of their time just navigating and engaging.  And people don't engage well with them, end up just hiding.  Unless there is a very clear need to move about, like oberserving some detailed experience,  like a store layout, you are better off just having a stationary video camera somewhere.   This might seem like a cute idea, but test carefully before you buy a fleet.

The Best Telepresence Robot
After spending 20 hours researching telepresence robots and testing two of the most promising models in office and home settings, we don’t think these devices are ready for prime time. But if you want a telepresence robot to give remote employees a physical presence in your office, the Suitable Technologies Beam Enhanced is the only bot that’s reliable and user-friendly enough to consider. .... " 

Friday, August 11, 2017

Data as Asset: Deriving Knowledge from Data

Good thought, and a natural way to lead toward value based intelligence.  In O'Reilly.

A DevOps approach to data management
A multi-model approach to transforming data from a liability to an asset.

By Adam Michael Wood 

Download the free O’Reilly report "Defining Data-Driven Software Development," by Eric Laquer.
Deriving knowledge from data has become a key competency for many—if not most—businesses. With the right data, and the right tools to handle it, businesses can gain keen insights into a variety of metrics, including operations, customer activity, and employee productivity. .... " 

Making AI Explain Itself

Back to our maintenance problem. Unless you can explain a result, it is hard to maintain it under changing context.   Trust too can be lost if the methods are indistinguishable from magic.  

Inside DARPA's Push to Make Artificial Intelligence Explain Itself 
The Wall Street Journal, Sara Castellanos; Steven Norton

The U.S. Defense Advanced Research Projects Agency (DAPRA) is coordinating a project in which 100 researchers at more than 30 universities and private institutions are seeking to create deep-learning artificial intelligences (AIs) that can explain their decision-making to humans. DARPA program manager David Gunning says this advance is crucial as AI becomes more deeply entrenched in everyday life and a greater level of trust between humans and machines must be nurtured. Participants have spent the project's first phase working on focus areas of their choosing, and in the second phase each institution will be assigned one of two "challenge problems" to address. The challenges will either involve using AI to classify events in multimedia, or training a simulated autonomous system to conduct a series of missions. The final result will be a set of machine-learning methods and user interfaces that public- or private-sector groups could use to construct their own explainable AI systems.  ... " 

Platt Research Institute: Journal of Retail Analytics

Have begun to follow, details for sign-in and article download at the link:

Journal of Retail Analytics

PRI’s Journal of Retail Analytics is a comprehensive quarterly publication that includes news and case studies regarding retail analytics, digital customer-facing technologies, and related topics. The Journal provides a snapshot of the economy as it impacts retailers and consumers. In addition, the Journal examines trends in the use of retail analytics and details developments in the digital communications industry. Authors include industry thought leaders and leading academics, among others.

PRI’s offices, projects, and extensive contacts in North America and Europe enable the firm to report on cutting-edge global events. Furthermore, the Journal features distinguished guest columnists, thought-provoking case studies, network profiles, and more.

If you are interested in submitting an article for publication in the Journal, please review the current Submission Guidelines.

The most recent edition of the Journal of Retail Analytics is available for free download below. Articles in the 2Q 2017 issue include:

Audio Management and Delivery Strategies for Retail Environments by Reto Brader, Vice President, Sales and Marketing, Barix

Building a Relationship That Online Shopping Can’t Replace by Richard Ventura, Vice President of Product Marketing and Solutions, NEC Display Solutions

Comparison of Traditional Predictive Analytics Tools Versus Artificial Intelligence-Based Solutions by Gary Saarenvirta, CEO, Daisy Intelligence

How Online Reviews Influence Sales by Spiegel Research Center, Northwestern University
Strategic Data Management in Retail Labeling Systems by Elizabeth Sinclair, Vertical Marketing Manager, Seagull Scientific

The Apple Store versus the Microsoft Store: Using Retail Analytics to Measure Customer Behavior Case Study — Part I by D. Anthony Miles, Founder and CEO, Miles Development Industries Corporation

The Role of Emerging Technologies in Retail: A Retail Roundtable   ... " 

The AI Nose

Have followed the idea of an 'Artificial Nose' for decades, we were first interested in using it as a automatic way to to evaluate and duplicate coffee blends and their components in the supply chain.  

Also, the argument was if we could digitize  such analyses, we could use that to deliver scents where we wanted to.  We did that in places like the retail store shelf and even the store aisle.   The solutions were never completely satisfactory.   The method here claims to use deep learning.  Continue to follow related tech, see the scent tag below.

An AI ‘nose’ can remember different scents
It can also detect potentially silent but deadly gas mixtures.    by Timothy J. Seppala, @timseppala  .... " 

Working in a Digitally Disrupted World

In McKinsey:

Do American workers feel they can ‘make it’?
Whether people in the United States believe they can thrive economically in a digitally disrupted world depends strongly on the amount of education they’ve attained, according to a new survey. ... " 

Thursday, August 10, 2017

How AI is Already Changing Business


An HBR Ideacast ...

How AI Is Already Changing Business

Erik Brynjolfsson, MIT Sloan School professor, explains how rapid advances in machine learning are presenting new opportunities for businesses. He breaks down how the technology works and what it can and can’t do (yet). He also discusses the potential impact of AI on the economy, how workforces will interact with it in the future, and suggests managers start experimenting now. Brynjolfsson is the co-author, with Andrew McAfee, of the HBR Big Idea article, “The Business of Artificial Intelligence.” They’re also the co-authors of the new book, Machine, Platform, Crowd: Harnessing Our Digital Future.

 Welcome to the HBR IdeaCast from Harvard Business Review. I’m Sarah Green Carmichael.
It’s a pretty sad photo when you look at it. A robot, just over a meter tall and shaped kind of like a pudgy rocket ship, laying on its side in a shallow pool in the courtyard of a Washington, D.C. office building. Workers – human ones – stand around, trying to figure out how to rescue it.

The security robot had just been on the job for a few days when the mishap occurred. One entrepreneur who works in the office complex wrote: “We were promised flying cars. Instead we got suicidal robots.”

For many people online, the snapshot symbolized something about the autonomous future that awaits. Robots are coming, and computers can do all kinds of new work for us. Cars can drive themselves. For some people this is exciting, but there is also clearly fear out there about dystopia. Tesla CEO Elon Musk calls artificial intelligence an existential threat.

But our guest on the show today is cautiously optimistic. He’s been watching how businesses are using artificial intelligence and how advances in machine learning will change how we work. Erik Brynjolfsson teaches at MIT Sloan School and runs the MIT Initiative on the Digital Economy. And he’s the co-author with Andrew McAfee of the new HBR article, “The Business of Artificial Intelligence.”

Erik, thanks for talking with the HBR IdeaCast.  ....    "

Marketing and AI

How will marketing adapt?
With AI, marketing is needed but marketers might not be
By Christine Coudert on SAS Voices ... " 

Robot Restocking

Includes a good expert discussion, usually the best parts of these pieces.   Also an approach we did much experimenting with, starting with Blackberry devices.

Can robots keep shelves stocked at Schnucks?  in Retailwire,  by George Anderson

Schnuck Markets, the 100-store supermarket chain based in St. Louis, announced it will run a six-week pilot program at three locations to test robots that will move up and down store aisles to make sure shelves remain properly stocked. The robots will also scan shelves to make sure that each item is in its proper place, aligned with the correct shelf tag.


The robots, which will be deployed three times a day (morning, afternoon and night), will send real-time information to store associates, helping them keep shelves stocked for customers. Each unit, named Tally by the manufacturer, Simbe Robotics, stands 38 inches high and weighs about 30 pounds. The devices use sensors to navigate around the store. The St. Louis Post-Dispatch reports that the units are programmed to avoid busy aisles and stop moving if a customer approaches.


Dave Steck, vice president of IT – infrastructure at Schnucks, said the data the robots collect will also be shared with vendors to help them improve their supply levels to stores. The initial emphasis of the pilot is to see if the chain can improve its stock positions through automation. Later, the technology may be used to reduce pricing errors and address other issues.  .... " 

Facebook Translates with Neural Nets

Another example  of how the power of this tech is expanding.  Some interesting details here.

Two versions of a Facebook post Facebook Translations Now Rely Entirely on Neural Networks in SiliconANGLE  by Eric David

" ... Facebook on Wednesday announced its translations are now wholly dependent on state-of-the-art machine-learning neural networks. A team of Facebook researchers says these networks manage more than 2,000 translation directions and 4.5 billion translations daily, generating more accurate translations than Facebook's previous system, which used phrase-based machine translation models. Neural machine translation gauges the complete content of a message together, which is more resource-intensive than phrase-based translation but typically results in a more fluent translation. Facebook also says neural machine translation can handle unknown or misspelled words with greater proficiency, as it can examine contextual clues to determine a word's intended meaning. Facebook thinks convolutional neural networks (CNN) can realize the same accuracy in translation as recurrent neural networks, but significantly faster. In May, the company announced that its CNN-based system was nine times faster than existing networks, and Facebook researchers note CNNs are a better fit for the newest machine-learning hardware. ... " 

Content Marketing in the Era of Voice-Controlled Devices

A number of interesting retail marketing examples are shown in this medium.

Content Marketing in the Era of Voice-Controlled Devices  in Flipboard
By Dawn Papandrea/Aug 9, 2017
Via John Frazier   BBA News - BizBuzz America

Voice-controlled devices are today’s hottest technology. 
Amazon Alexa now has more than 15,000 skills (which can range from basic functions to more robust experiences, akin to apps.) And each day, we’re hearing about new capabilities or big announcements for similar devices. Apple, for one, will release a Siri-based HomePod this fall. 

As a content marketer, you should be keeping your ears (and eyes) open to these new audio channels to ensure that your brand voice is heard. 

“If a brand doesn’t have a presence on voice platforms, then they are literally silent when a consumer asks to engage with them,” says Bret Kinsella, Editor and Publisher of Voicebot.ai, a publication about the voice and AI revolution.  

Are you ready to be an early adopter and develop audio content to expand your reach? If so, listen up: Here’s what you need to know about the voice-controlled device space.  .... "

Wednesday, August 09, 2017

Data Fracking Dark Data

Found this to be an interesting,  the idea of 'Data Fracking' ... a means of finding the right data.   Better way to find and use your data assets?  Improving access to the right data.  Broadly using the oil drilling method as metaphor.  Developed by a company called Datumize.  Through the use of the right Dark Data, which is unused data in or outside your company.

" .... Effectively exploiting this resource will require new techniques.  Data Fracking™ is a new approach focused on the discovery, collection, integration and deployment of Dark Data so it can be collected, refined and made available for use to enhance your operational and decision making processes.  Data Fracking™ enables the discovery, collection, integration and utilization of this previously untapped resource. .... " ,   Have not tried this as yet, but plan to follow up.  Piece on this in DSC.   And a white paper on the method.
  

How Can Humans Keep Control of AI?

This starts with a clear understanding of what AI's, or anything that appears to be autonomous in any way,  is operating, and why it is operating as it does.  This has happened since the dawn of computing.   What is new is the breadth and depth of possible autonomy  Plus its transparency has decreased, this mostly in the last few decades.   Not enough to be dangerous yet today,  but the direction in which this is moving needs to be watched.   We need to understand this more than ever before.

Related, see also more from the Leverhulme Center on AI and the Future of Humanity, tag below.

How Can Humans Keep Control of AI? 
The Nation (Thailand Portal)  Via CACM      August 8, 2017

In an interview, Osaka University president Shojiro Nishio stresses the need to keep artificial intelligence (AI) under human control. Nishio notes a key issue is a lack of human understanding of how an AI arrives at certain answers, and he says a major area for researchers is improving how to check all processes the AI went through. Nishio also says this "black box" problem demands an international response, given that AI will eventually be connected to the Internet and have a global impact. Nishio cites a draft plan from Japan's Internal Affairs and Communications Ministry directing developers to ensure they can explain the results of their AI's judgments, and that they can control their AI. "I am more concerned about what happens if AI produces incorrect answers or develops in a way that humans cannot control," Nishio notes. "That sort of thing happens more and more often as AI starts getting used to real-life situations." .... 

How can humans keep control of AI   By The Yomiuri Shimbun 
Asia News Network , Tokyo

Expectations for artificial intelligence are expanding, and people are making more and more attempts to use AI in a variety of areas, from finance and medical care to manufacturing.
However, some are sounding warning bells and say that AI will eventually exceed the capabilities of humans. Shojiro Nishio, president of Osaka University, has been working on computer research for many years.  He speaks on AI-related arguments and issues we can expect in the future.  .... " 

Essence of Voice Interfaces

Very good piece here, what is the difference and essence of having voice interfaces?  Or even interfaces that are partially voice?  This is not necessarily AI but it is closer to cognitive than we had before.  Remember too, that voice more strongly implies a conversation is occurring, which embeds, context, memory and modeling who/what you are talking to.  None of this has been perfected.   Clear from my interaction with Echos and Google Home during the past few years.

“Alexa, Understand Me”  In Technology Review  by George Anders
Voice-based AI devices aren’t just jukeboxes with attitude. They could become the primary way we interact with our machines.

On August 31, 2012, four Amazon engineers filed the fundamental patent for what ultimately became Alexa, an artificial intelligence system designed to engage with one of the world’s biggest and most tangled data sets: human speech. The engineers needed just 11 words and a simple diagram to describe how it would work. A male user in a quiet room says: “Please play ‘Let It Be,’ by the Beatles.” A small tabletop machine replies: “No problem, John,” and begins playing the requested song.

From that modest start, voice-based AI for the home has become a big business for Amazon and, increasingly, a strategic battleground with its technology rivals. Google, Apple, Samsung, and Microsoft are each putting thousands of researchers and business specialists to work trying to create irresistible versions of easy-to-use devices that we can talk with. “Until now, all of us have bent to accommodate tech, in terms of typing, tapping, or swiping. Now the new user interfaces are bending to us,” observes Ahmed Bouzid, the chief executive officer of Witlingo, which builds voice-driven apps of all sorts for banks, universities, law firms, and others.  .... " 

Versustech

Recenty looked at efficiency delivery in healthcare systems.  Here is one company that links location to efficiency.   Versustech.    Real time location system  (RTLS)  a part of Midmark.

Amazon Echo in Physical Store

Amazon Echo used in Physical Store for Customer Engagement

In LSN  Global: 

4. Tuft & Needle embraces Amazon
Seattle – The mattress start-up, which reportedly gets 25% of its sales through Amazon, will open its first physical store in the mega-system’s home town later this year.

The space will feature tablets that enable visitors to read Amazon reviews, Amazon Echo devices to answer customers’ questions and QR codes on products to enable seamless payments through the Amazon app.  ... "

Coursera on Deep Learning with Andrew Ng

Build Your Career in AI

Take our new Deep Learning  courses, now open on Coursera  

Enroll.

By Andrew Ng

More on the courses.

Tuesday, August 08, 2017

Free Deep Learning Book Completed

Data Science Central points to a free book on deep learning by the MIT press.   Ultimately very technical, but the introductions are useful for anyone interested in the topic.   Table of contents and links to all sections at the link below.

Free Deep Learning Book (MIT Press)  Posted by Vincent Granville  
The Deep Learning textbook is a resource intended to help students and practitioners enter the field of machine learning in general and deep learning in particular. The online version of the book is now complete and will remain available online for free. .... " 

This is the same book that was mentioned previously, it is now completed.

Pop-up Stores to Tell CPG Brand Stories

Was involved in several such efforts, in CPG, but not food.   Brands were new.  The approaches were first mocked up in laboratory settings.  Did not see specific financial outcome estimates.  More discussion at the link

Why is big food turning to pop-up stores to tell brand stories?  by Dale Buss in Retailwire
"A pop-up store allows the brand to shine and promote its best qualities without the competition breathing down its neck."   Zel Bianco, President, founder and CEO Interactive Edge

Through a special arrangement, presented here for discussion is a summary of a current article from the monthly e-zine, CPGmatters. ... 

Kellogg, Pure Leaf and Chobani are among a number of CPG companies that have been sprouting “pop-up” stores that tell their brand stories. As well as fashioning “stores within a store” in larger retailers, CPG brands are putting these pop-ups into semi-permanent locations on street fronts in New York City and other heavily populated locales.  ... " 

IBM Research Writes about Progress in Deep Learning

Recently have used deep learning from IBM to solve a difficult classification problem.

From IBM AI Research: 
New IBM Research distributed deep learning software achieves record performance for large neural network, large data set ... 

Deep learning is a widely used AI method to help computers understand and extract meaning from images and sounds and other data types using neural networks, a brain-inspired approach to computing. It holds promise to fuel breakthroughs in everything from consumer mobile app experiences to medical imaging diagnostics. Progress in accuracy and deploying deep learning at scale is limited by technical challenges that slow processing time to days and weeks. IBM Research AI experts have created distributed deep learning software, achieving record performance for image recognition accuracy and large neural networks composed of up to 250 GPUs, a special processor for large amounts of data. Developers and data scientists can now preview this technical milestone in version 4 of the PowerAI enterprise deep learning software.  ... " 

Assistants Addressing Primary Health Care

Primary Health care via assistant, continue to follow this topic.   Good overview:

The robot will see you now  By Pam Baker in HP Enterprise
Intelligent personal medical devices will improve your quality of life. And nag you when you have that extra dessert.

While everyone in the U.S. is focused on health insurance, the truth remains that there are not enough primary care doctors to tend to everyone. Regulation (or deregulation) alone can’t fix that shortage. In order to provide prevention care, illness and injury treatments, and life-saving measures for all, technology must fill in and assist human healthcare providers.

But this is not a situation unique to the U.S., as all countries around the globe struggle with a shortage of healthcare professionals and resources in varying degrees. Further, healthcare providers also struggle with work overloads, unacceptable error rates, and lost revenue from complexities and errors in payer claims.  .... " 

Raise Your Hand With Glitch

Nice simple idea, not sure how rewards work, but could see this working among an enterprise team and getting some sort of reward points for helping. Could also push toward standard usage when developing.  Ultimately might be a way to integrate bot assistance.

Just raise your hand: How Glitch helps

With our brand new upgrade to Glitch, getting coding help in real time is as easy as just raising your hand.

Let’s face it: when we’re programming, we’re all beginners sometimes. Whether we’re brand-new coders learning the basics of how to build an app, or seasoned professionals trying out new APIs and web services, part of being a programmer is putting yourself in unfamiliar territory.

But trying new things can be a little intimidating. So we’ve invented a brand new way to get help in realtime with Glitch: it’s as easy as raising your hand. .... " 

Global Value Chains

Internalizing Global Value Chains: A Firm-Level Analysis  in HBS
by Laura Alfaro, Pol Antràs, Davin Chor and Paola Conconi

In recent decades, advances in information and communication technology and falling trade barriers have led firms to retain within their boundaries and in their domestic economies only a subset of their production stages. A key decision facing firms worldwide is the extent of control to exert over the different segments of their production processes. We describe a property-rights model of firm boundary choices along the value chain that generalizes Antràs and Chor (2013). To assess the evidence, we construct firm-level measures of the upstreamness of integrated and nonintegrated inputs by combining information on the production activities of firms operating in more than 100 countries with input-output tables. In line with the model's predictions, we find that whether a firm integrates upstream or downstream suppliers depends crucially on the elasticity of demand for its final product. Moreover, a firm's propensity to integrate a given stage of the value chain is shaped by the relative contractibility of the stages located upstream versus downstream from that stage, as well as by the firm's productivity. Our results suggest that contractual frictions play an important role in shaping the integration choices of firms around the world.    ... " 

Monday, August 07, 2017

Patent Troll Loses Podcast Fight

Was slightly involved with research for patent litigation that dealt with trolls,  so follow recent activity in this area, in particular regarding prior art.    Quite a bit of detail at the link regarding podcasting prior art.  The patent system has to be fixed.

Personal Audio loses its appeal for podcasting patent   by Brian Heater
A year after taking up the case, the US Court of Appeals for the Federal Circuit has ruled in favor of the Electronic Frontier Foundation in its challenge against podcasting patent troll, Personal Audio. The decision is a massive relief for the vibrant and ever-growing medium, which has been operated under the threat of lawsuit for a number of years.  .... " 

Siemens: Getting Robots and Workers to Cooperate

Like that this approach is addressing elements and architecture of process, and that the design and operation of  these elements is also a big part of doing something intelligently.  New component elements of business process modeling?   Looking deeper.

This AI Factory Boss Tells Robots And Humans How To Work Together  in FastCompany.

Siemens’s future “artificially intelligent” factory is aimed at a new trend in automation: telling robot and human workers how to help each other make things.

Robots are consistent, indefatigable workers, but they don’t improvise well. Changes on the assembly line require painstaking reprogramming by humans, making it hard to switch up what a factory produces. Now researchers at German industrial giant Siemens say they have a solution: a factory that uses AI to orchestrate the factory of the future, by both programming factory robots and handing out assignments to the humans working alongside them.

“Instead of programming [each robot], what we do is say, this is a machine that can do this [task], this is a machine that has the following capabilities,” says Florian Michahelles, who heads the Siemens Web of Things research group in Berkeley, CA. An AI program that Michahelles and his team have developed, called a “reasoner,” figures out the steps required to make a product, such as a chair; then it divides the assignments among machines based their capabilities, like how far a robotic arm can reach or how much weight it can lift.  ...  "

Future of Sales is AI

Looking at the future of sales.   Further, if you can automate it, you can improve it.  In performance, reach, cost or adaptability.  And learn about some of these same contexts.   Not to say this is new, we have done these things for a long time.  What is new is the introduction of cognitive channels to people we never had before.  That is what AI is in the short term.  Not thinking machines, but machines that help us by augmenting our thinking in familiar ways.

The Future Of Sales Is In Artificial Intelligence, According To This CEO
If you can automate it, you can scale it. By Nicolas Cole,    In Inc.   .... " 

Its All About Continuous Learning

Our own early experiments in AI failed because we ended up needing expensive continuous Maintenance.   We need continual learning, and we need to plan for that up front.  I always recommend that for projects that seek to integrate intelligence in process.   Feedback will likely be sparse at first,  but needs to be built-in and well designed.

Good piece in O'Reilly:

Why continuous learning is key to AI
A look ahead at the tools and methods for learning from sparse feedback.   By Ben Lorica August 7, 2017

" ... I take for granted that future AI systems will rely on continuous learning as opposed to algorithms that are trained offline. Humans learn this way, and AI systems will increasingly have the capacity to do the same. Imagine visiting an office for the first time and tripping over an obstacle. The very next time you visit that scene—perhaps just a few minutes later—you’ll most likely know to look out for the object that tripped you. .... " 

Voices in AI Podcasts

From GigaOM, have signed up for these.

Announcing “Voices in AI” Podcast,  Byron Reese
Artificial intelligence has received more digital ink recently than just about any topic in technology. This level of coverage is warranted, since AI reaches new milestones seemingly every week. So whenever AI exceeds human capabilities in some new area, we collectively reflect on what this all means and speculate about what the machines might do next. And on this topic, everyone has opinions, including Elon Musk, Mark Zuckerberg, Bill Gates, Mark Cuban and Stephen Hawking.

The interesting thing to me is how different all of their opinions are. Some think the AIs will take away our jobs; others believe they will create better ones for everyone. Some say we should fear the AIs while others maintain that such concerns are science fiction nonsense. Some predict we will get a general AI in a few years; others think a few centuries. ... " 

Sunday, August 06, 2017

Google's Gmail will Answer Your Emails?

What is Ray Kurzweil up to at Google?  Writing your Emails.  ... in Wired ..  I respond:

Interesting background and information about an innovator and his team's neural net driven AI project.  Good as an indication of how Google works internally with innovative ideas.  Worth readng for that.  Kurzweil is very well known, innovative and very smart.

But here is a case where I have been actually 'using' the results of his team's work since it was made public on Google Mail in May.   I would say I use one of his auto answers or 'Smart Replies'  maybe one out of twenty times.  Then worry that the recipient might notice.   Have never edited an answer, but understood that I could.  Just easier to write the whole thing.

  And must say that I am not overly impressed.  This could have been done better with a lot less resources and effort.  I would not pay for this if I had to, not even close.   Is trite a good word?   Not what I expected when I heard it called  'automatic answering'.   The latter is really hard, at least to do well.   I would have to read it very closely to see if it is correct.  If its not that important, why answer at all?    More and better is coming they say, I hope so.   My singularity receptacle is waiting. But you will have to catch me.

Conversational Banking Assistant Pilot

Intriguing statement of how this is operating, and how it can deal with a messy conversational interaction.   Because of the nature of the interaction, this would appear to be particularly important and difficult to assure correctness.   How exactly is this beyond rules or a chatbot interaction?   Sentiment also included.  Learning too.    More detail at the link.  ...

USAA Rolls Out Innovative Conversational AI Solution
Going beyond traditional rules-based voice or chatbot digital banking solutions, a non-bot, natural language banking experience is being offered to USAA members in an Amazon Alexa pilot with Clinc.    By Jim Marous, Co-Publisher of The Financial Brand and Owner/Publisher of the Digital Banking Report

Subscribe TodayLast September, a team of computer science professors at the University of Michigan introduced an application developed in their research lab that they believed would change the way consumers would do banking in the future. Combining the science and technology from academia, with the needs for a better voice-first mobile banking capability, Clinc won ‘Best of Show’ honors at Finovate in New York City with their Finie (“the financial genie”) application.

The intelligent personal assistant uses sophisticated natural language processing engines that have been trained with a deeper knowledge of the financial and banking industry as opposed to using a rules-based approach. Unlike solutions that currently exist from Siri, Alexa and Cortana, Clinc’s machine learning capability allows the application to expand knowledge and improve responses with every query. By identifying and analyzing dozens of factors like speech patterns, word structure and sentiment, Clinc is able to understand, to remember and to respond to unconstrained, contextual, messy human language.  ... " 

Transcript Data to Tailor Voice Interaction

Amazon’s Alexa Has A Data Dilemma: Be More Like Apple Or Google?
Amazon’s virtual assistant is caught between Apple-like privacy safeguards and Google-style openness.   By Jared Newman

Devices like Amazon Echo could someday turn into a treasure trove for developers that make voice assistant skills, but first companies have to figure out where they draw the line when it comes to weighing data sharing against consumer privacy.

Now that dilemma is heating up: Citing three unnamed sources, The Information reported this week that Amazon is considering whether to provide full conversation transcripts to Alexa developers. This would be a major change from Amazon’s current policy in which the company only provides basic information—such as the total number of users, the average number of actions they’ve performed, and rates of success or failure for voice commands. Amazon declined to comment to The Information regarding the claims, but the change wouldn’t be unprecedented. Google’s voice assistant platform already provides full transcripts to developers.  ... " 

Detecting Randomness of Numbers

Early projects had us looking to determine if numbers were random or not.    Signal vs noise detection.   Can be very important for scrubbing data before use.   Here an example of a means of detecting non-randomness.   In DSC. 

Saturday, August 05, 2017

Watson Machine Learning Now Publicly Available

I note that this also includes 'visual' machine learning,  using methods that look like the Clementine system they acquired long ago for System Modeler.   Good direction.   Explored some of this via their Bluemix services, but this takes it further, for both professionals and for non professional use.

I contractually examined several examples of enterprise use in real projects, impressive plug-in power.  Natural links to Watson Analytics?  Looking at the new usability here and cost.

Watson Machine Learning is now Generally Available
Today we are excited to announce the general availability of the IBM Watson Machine Learning service. Over the past 12 months we've got feedback from hundreds of beta users of the Watson Machine Learning (WML) service. During the beta period, we’ve been actively collecting feedback provided via email, Slack, and targeted surveys. The WML product team has been actively engaged in those conversations and wherever possible we’ve worked to incorporate your feedback in to the service. With today’s announcement, we are now opening this service to the general public and rolling out a number of new features. Read on to learn more…

What is WML and why are we building it?
WML is a Bluemix service that enables users to perform two fundamental operations of machine learning. .... 

Training: this is the process of refining an algorithm so that it can 'learn' from a dataset. The output of this operation is called a model. A model encompasses the learned coefficients of mathematical expressions.

Scoring: the operation of predicting an outcome using a trained model. The output of the scoring operation is another dataset containing predicted values.   ... " 

More details.


Automated Machine Learning Tools

William Vorhies surveys some 'automated' data machine learning systems.  Agree these are for professionals,  but can likely decrease the effort needed to produce such models.  But their emergence can only mean that these techniques will ultimately be automated more fully. ...

Automated Machine Learning for Professionals
Posted by William Vorhies  

Summary:  There are a variety of new Automated Machine Learning (AML) platforms emerging that led us recently to ask if we’d be automated and unemployed any time soon.  In this article we’ll cover the “Professional AML tools”.  They require that you be fluent in R or Python which means that Citizen Data Scientists won’t be using them.  They also significantly enhance productivity and reduce the redundant and tedious work that’s part of model building.   ...   "