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

Tuesday, January 31, 2017

AI Newsletter

O'Reilly's latest AI Newsletter.  Usually informative.  Often linking to their books and related materials and webcasts.  Worth a follow.

Why Should Algorithms Explain themselves?

Some of our earliest work in delivering expertise via AI to the enterprise failed because the system could not explain its reasoning to its users.   The system could not sell itself.   But does a system need to do that?  Or does it just have to be right most of the time?   Or right enough for the business purpose?  Search, as it has trained us, gives us a list of results, some irrelevant, but we have become used to choosing the best. And usually go away satisfied.  So is explanation even necessary? Or if explanation is not needed,  is it still very useful?

In QZ:  Algorithms explaining themselves.

MS Buys Deep Learning Company

Continuing to be interested in what 'conversation' means for AI.    For humans it is a key part of sharing, delivering, learning and developing expertise.  How do we give that to machines?

Deep Learning Startup Maluuba Acquired by Microsoft, AI Expert Yoshua Bengio to Become Advisor
Microsoft's acquisition of startup Maluuba will help with the ultimate goal of a thinking, conversational AI. The exact role of the acquisition is not yet clear, but the Redmond giant will be sharing more 'in the coming months.' ....    By Ryan Maskell .... 

Apple Making Smart Glasses

Now back to the idea of augmented reality glasses, where Google glass failed, though is apparently still being worked.  Do we want fashion, or actual value, or both?    Which drives sales the most, adoption the most?

Why Apple will make smart glasses
Instead of making unfashionable smart glasses, Apple will make fashionable glasses smart. ... " 

Salespeople and Executives

Worked on executive information systems that linked sales people and executive decisions.

Executives and Salespeople Are Misaligned — and the Effects Are Costly
Frank V. Cespedes, Christopher Wallace

U.S. companies spend over $900 billion on their sales forces, which is three times more than they spend on all ad media. Sales is, by far, the most expensive part of strategy execution for most firms. Yet, on average, companies deliver only 50% to 60% of the financial performance that their strategies and sales forecasts have promised. And more than half of executives (56%) say that their biggest challenge is ensuring that their daily decisions about strategy and resource allocation are in alignment with their companies’ strategies. That’s a lot of wasted money and effort. .... "  

Monday, January 30, 2017

Smart Glasses Adjust for Visual Tasks

Not augmented reality glasses, but those that adapt to needed vision.

I Can See Clearly Now
UNews (UT) (01/24/17) Vincent Horiuchi 

University of Utah researchers have created "smart glasses" with liquid-based lenses that can automatically adjust focus on what a person is seeing, whether it is far away or nearby. The smart eyeglass lenses are made of glycerin, a thick colorless liquid enclosed by flexible rubber-like membranes in the front and back. The rear membrane in each lens is connected to a series of mechanical actuators that push the membrane back and forth like a transparent piston, changing the curvature of the liquid lens and the focal length between the lens and the eye. "The focal length of the glasses depends on the shape of the lens, so to change the optical power we actually have to change the membrane shape," says Utah professor Carlos Mastrangelo. He notes the lenses are placed in specialized eyeglass frames equipped with electronics and a battery to control and power the actuators. The bridge of the glasses has a distance meter that measures the distance from the glasses to an object via pulses of infrared light. When the user looks at an object, the meter instantly measures the distance and automatically adjusts the lenses. If the user sees another, closer object, the distance meter readjusts and tells the actuators to reshape the lens.  .... " 

What is Machine Learning, and Why is it Intelligent?

Not bad, completely non-technical, 5 minute, video introduction to machine learning.  And why it is a kind of 'intelligence'.  Nicely done.

Ankit Mistry
Senior Software Engineer at Citibank India

What AI Needs is a Dose of Realism

Good thoughts, though I am a strong proponent of AI, there is still much to do.

What AI needs is a dose of realism
Oren Etzioni talks about the current AI landscape, projects at the Allen Institute, and why we need AI to make sense of AI.

Video by Jenn Webb January 5, 2017  ,,, 

"I think what's missing in the AI conversation is a dose of realism. We have on the one extreme people like Kurzweil, who are fantastically optimistic but don't really have data to back up their wildly optimistic predictions. On the other hand, we have people who are very afraid, like Nick Bostrom, who's a philosopher from Oxford, or Elon Musk, who needs no introduction, and say AI is like summoning the demon, which is really religious imagery. But again, neither party has the data to base their conclusions on; it's wild extrapolations, it's metaphor (like AI is a demon), it's philosophical argumentation.

"I think we need to have a more measured approach, where we measure AI's performance, where we understand that superhuman success on a narrow task like Go doesn't translate to even human performance level on a broad range of tasks, the kind that people do. I like saying my six-year-old is a lot smarter than AlphaGo—he can cross the street, more or less."

—Oren Etzioni

Year of the Chatbot?

Is it the year of the chatbot? Many examples out there, but still needs considerable evolution.  May well be the year of chatbot experimentation.

Year of the Chatbot?   

Through a special arrangement, presented here for discussion is a summary of a current article from the Retail TouchPoints website.

If 2016 was the year chatbots entered the retail scene, 2017 will be the year brands will start to realize real business results from the technology, through increased sales, conversion rates or customer loyalty.

While the overall rate of adoption among retailers is still nascent, Nordstrom, 1-800-Flowers, American Eagle Outfitters and Sephora were among those implementing such intelligent assistance tools on messaging platforms such as Facebook Messenger and Kik. Leveraging consumer data with artificial intelligence enabled the retailers to perform a wide range of tasks, including  ...  " 

Differential Event Models

Having been a long time practitioner of simulation using events, this immediately attracted me  to this description.   Technical.  Thinking how this can be used, have not completely understood it yet.   Interesting too are the kinds of applications mentioned.

Crosswave Differential Event Model    Posted by Don Philip Faithful
I have been writing about the Crosswave Differential Algorithm for a number of years. I described in previous blogs how the algorithm emerged almost by accident while I was attempting to write an application intended to support quality control. In this blog I will be discussing the event model that powers the algorithm. Events are the details and circumstances surrounding the numbers used in calculations. A metric such as revenues might be disconnected from the ideas and circumstances that brought about purchasing decisions. Events provide the story behind the numbers. However, there are an infinitesimal number of events. It is necessary to recognize those events might later be found relevant. An "event model" is the means by which event selection occurs. .... " 

Random Walks for Time Series Forecasting

Nicely done, fairly gentle.

A Gentle Introduction to the Random Walk for Times Series Forecasting with Python   by Jason Brownlee on January 20, 2017 in Time Series

How do you know if your time series problem is predictable?

This is a difficult question with time series forecasting. There is a tool called a random walk that can help you understand the predictability of your time series forecast problem.

In this tutorial, you will discover the random walk and its properties in Python.
After completing this tutorial, you will know:

What the random walk is and how to create one from scratch in Python.
How to analyze the properties of a random walk and recognize when a time series is and is not a random walk.
How to make predictions for a random walk.  .... " 

Sunday, January 29, 2017

Google Testing D-Wave Quantum Computing

A recent article led me to this arxiv.org paper abstract.  (Technical)  Using the D-Wave hardware.  Are the methods being shown real quantum computing?   Points to how Google is testing it.  This relates to some of our own experimentation.  Its often useful for finding global approximate optimums.  I do note the caution about global interactions and real-world problems:   " ... A recent Google study [Phys. Rev. X, 6:031015 (2016)] compared a D-Wave 2X quantum processing unit (QPU) to two classical Monte Carlo algorithms: simulated annealing (SA) and quantum Monte Carlo (QMC). The study showed the D-Wave 2X to be up to 100 million times faster than the classical algorithms. The Google inputs are designed to demonstrate the value of collective multiqubit tunneling, a resource that is available to D-Wave QPUs but not to simulated annealing. But the computational hardness in these inputs is highly localized in gadgets, with only a small amount of complexity coming from global interactions, meaning that the relevance to real-world problems is limited.  ... "  

Watson Analytics Professional

Watson Analytics Professional: 3 reasons to go pro   by Karen Galiger

Watson Analytics Professional Edition was just the right solution for Mears Group, LLC. With the heavy responsibility of servicing and maintaining 1 million homes in the UK, they needed help. They analyzed data with Watson Analytics and found new insights on targeting services. Now they respond more rapidly to emergencies. And, they can better plan maintenance. Watson Analytics Professional was just right for their big job.

Of course, all Watson Analytics editions make it easy for you to find insights on your own. You ask, discover, predict and visualize your way. The result? Better business understanding and smarter decisions. But, Watson Analytics Professional gives you the power of supersized easy analytics. .... " 

Learning and Bayesian updating in long cycle made-to-order production

My colleague Jeff Camm writes in a recent paper. The Bayesian learning aspect of parameters is interesting.  Technical.  Abstract.

Learning and Bayesian updating in long cycle made-to-order(MTO) production
K. Womer, H.Li, J. Camm, C. Osterman, R. Radhakrishnan ... 

We model production planning for made-to-order (MTO) manufacturing by choosing production rate to minimize expected discounted cost incurred up to a promised delivery date. Products that are MTO are often unique and customized. The associated learning curve slope and other production parameters cannot be precisely estimated before production starts. In this paper, a dynamic and adaptive approach to estimate the effects of learning and to optimize next period production is developed. This approach offers a closed-loop solution through stochastic dynamic programming. Monthly production data are used to update the joint probability distributions of production parameters via Bayesian methods. Our approach is illustrated using historical earned-value data from the Black Hawk Helicopter Program. Managerial insights are obtained and discussed. ... " 

Saturday, January 28, 2017

Let Alexa Scan Paper Abstracts for You

Alexa gets Meta, finds you the latest AI research
The Echo bot can find the 50 latest papers and read you the summaries.
Steve Dent, @stevetdent  in Endgadget. 

I examined this.  Simply and nicely done.  If you want to scan the abstracts of newly published papers in your field, here is a way to do it. The domain example used is Machine Learning, the skill is called arxvML, the full code is published for free, and I imagine could be adapted for any domain with some work.    Probably best for scanning abstracts, not sure if Alexa will do a good job of reading equations to you, or you will understand them spoken.

   Also likely works where you have to get alerted to many things, constantly changing, hands or full attention-free.  As in most skills, the verbal interface is a bit rocky, needs to be learned, and implies that what you are doing is very valuable or time-saving.    Could see this done for other kinds of segmented and context intelligent podcasts?   Thinking of other examples.

Detecting and Identifying Images

Home/News/Air Force Tests IBM's Brain-Inspired Chip as an Aerial Tank Spotter

ACM Tech News
Air Force Tests IBM's Brain-Inspired Chip as an Aerial Tank Spotter.   Neuromorphic chips.
In Technology Review

Beyond Big Data to Complex Systems

Well written challenge to when big data doesn't work, and an intro to complex systems.

When Big Data Isn’t Enough    Alex Woodie

The big data paradigm has changed how we make decisions. Armed with sophisticated machine learning and deep learning algorithms that can identify correlations hidden within huge data sets, big data has given us a powerful new tool to predict the future with uncanny accuracy and disrupt entire industries. But what if data alone isn’t enough? What if some decisions can’t be based just on data?

That may sound like a heresy to people who have devoted themselves to the religion of data, to the business leaders who have declared their allegiance to making data-driven decisions. Sure, there may be obstacles to overcome, such as data cleanliness, governance, and security concerns. But armed with enough data, one can solve any challenge, they say. Just look at the results. Right?  ... " 

Smart Machines to Boost Emotional Intelligence

Intriguing premise.  As an engineer I don't think I fully understand it.  Link to podcast.

Author Edward Hess discusses his new book, Humility is the New Smart: Rethinking Human Excellence in the Smart Machine Age ....   In Knowledge@Wharton:

"... Technology in the so-called Smart Machine Age, which includes AI, virtual reality and robotics, will bring huge changes not just in headcount, but also in how people innovate and collaborate. That will require new approaches to how people think, listen and relate, says Edward D. Hess, a professor of business administration at the University of Virginia. In the “Smart Age” now evolving, ego has no place. Instead, the focus will need to be on the quality of ideas, accuracy, emotional intelligence and mindfulness. Hess writes about these issues in the just-released book he co-authored with Katherine Ludwig, titled Humility is the New Smart: Rethinking Human Excellence in the Smart Machine Age. Hess discussed his ideas on the Knowledge@Wharton show on Wharton Business Radio, SiriusXM channel 111. (Listen to the podcast at the top of this page.) .... " 

Friday, January 27, 2017

Mr Clean Goes to the Superbowl

Have been a participant in the Mr Clean brand equity, and especially how it has been enhanced by technology.    Lately there has not been much technical enhancement, but at least he is going to the US Superbowl.  He has also learned some personal cleaning skills of his own.   In AdAge.

Virtualitics for VR Analytics

Newly brought to my attention:  Virtualitics.  Bringing VR to Analytics. Another thing we experimented with.  More specifics and tests will follow.  In Beta:

" ... We have built the first platform to merge Artificial Intelligence, Big Data and Virtual/Augmented Reality.

Next generation data analytics ... Visualize and understand your data as never before thanks to our innovative use of VR/AR, Machine Learning and Natural Language.

Shared Virtual Office (SVO)
Collaborative and fully customizable shared space where you can analyze data and present and discuss insights.

Fast, Reliable, Smart, Secure
Complete your tasks in a tenth of the time needed with traditional software and take advantage of smart routines.

Machine Learning and Natural Language
Smart mapping based on Machine Learning and Natural Language that summarizes the key patterns in the data. ..... " 

Getting Better Wisdom from Crowds

Like the quantifying aspect of this.  We experimented with WOC choices and how they linked with actual commercial decisions.   Would like to see it tested with  consumer as well as expert groups.

Better wisdom from crowds
MIT scholars produce new method of harvesting correct answers from groups.
Peter Dizikes | MIT News Office 

The wisdom of crowds is not always perfect. But two scholars at MIT’s Sloan Neuroeconomics Lab, along with a colleague at Princeton University, have found a way to make it better.

Their method, explained in a newly published paper, uses a technique the researchers call the “surprisingly popular” algorithm to better extract correct answers from large groups of people. As such, it could refine wisdom-of-crowds surveys, which are used in political and economic forecasting, as well as many other collective activities, from pricing artworks to grading scientific research proposals.

The new method is simple. For a given question, people are asked two things: What they think the right answer is, and what they think popular opinion will be. The variation between the two aggregate responses indicates the correct answer.

“In situations where there is enough information in the crowd to determine the correct answer to a question, that answer will be the one [that] most outperforms expectations,” says paper co-author Drazen Prelec, a professor at the MIT Sloan School of Management as well as the Department of Economics and the Department of Brain and Cognitive Sciences. .... " 

Allowing End Users to Build Models

Very good general question that we have been discussing for years, well, really decades.   Started way back in the spreadsheet era, when you empowered users to build 'models' of any kind.    Afterwards we were sometimes brought in to clean up the mess.     I would go back to a risk analysis.  What are the risk implications of giving a wrong answer?  How wrong?  Depending on the answer to that, what kind of model checks the answer?  Who/what decides the answer is right?  Need the results be formally checked at all? How can this cautionary checking be properly automated?

Discussion in Linkedin:
Is it risky to let users create their own predictive models?    by Michael Surkan

Adding AI to Raspberry Pi

Could be quite  a way to add AI experience for students.  In CWorld: 

Google wants to add A.I. to Raspberry Pi gadgets
Google plans to bring open source machine-learning tools to board makers later this year
By Agam Shah .... "

Thursday, January 26, 2017

Q&A on PowerAI

For research I took a look at the PowerAI offering.  See the complete article:

Putting the “AI” in PowerAI
A Q&A with IBM Research’s Hillery Hunter

IBM’s latest Power servers come with an AI twist. Optimized for deep learning, a new so-called PowerAI toolkit will “help train the systems to think and learn in a more human-like way, at a faster pace,” as announced at SC16, the International Conference for High Performance Computing, Networking, Storage and Analysis.

I spoke with Hillery Hunter, IBM Research’s director of Systems Acceleration and Memory and Memory Strategist, about her team’s contribution to the software behind the world’s fastest deep learning servers.

How was the AI added, and what does that mean for how a Power system functions?
 Today’s launch represents our first public offering in hardware-software co-optimization for deep learning. Researchers have worked closely with IBM’s systems engineers to create code that is optimized to the Power S822LC (IBM’s highest-performing OpenPOWER server).

We used the S822LC’s unique design – including high-bandwidth NVLink interconnect, not just between pairs of GPUs, but also between the GPUs and the CPUs – to deliver higher deep learning training performance.

This means that the co-designed PowerAI hardware and software can build learned models from images, speech, or other media in less time than prior generations of hardware and software. Deep learning training time is a key metric for developer productivity in this domain. It enables innovation at a faster pace, as developers can invent and try out many new models, parameter settings, and data sets.    .... " 

Tensorflow and PowerAI

Great to see integration of key AI software, lets also start to see excellent examples of their use with supporting data architecture.

IBM brings Google's AI tools to its Power hardware

IBM is integrating the open-souce TensorFlow into its PowerAI toolkit
By Agam Shah  

Google has cool technology to recognize images and speech, and IBM's hardware can diagnose diseases. ... Combine the two, and you get a powerful computer with serious brains.

IBM is merging Google's artificial intelligence tools with its own cognitive computing technologies, allowing deep-learning systems to more accurately find answers to complex questions or recognize images or voices.

Google's open-source TensorFlow machine-learning tools are being packed into IBM's PowerAI, which is a toolkit for computer learning. The two can be combined to improve machine learning on IBM's Power servers.

A computer learns as more data fed into its system, much like how a human learns. PowerAI and TensorFlow tools can help track patterns and classify data, and spit out approximate answers to queries. The answers will be more accurate as a computer learns more.  .... " 

Talk on Technology Trends

William Chamberlin from IBM Market Development and Insights, presented "IT Technology Trends in 2017... and Beyond."     Slides.   Talk.

 Use the discussion group https://www.linkedin.com/groups/Cognitive-Systems-Institute-6729452  to discuss further.

Update on D-Wave

Been following, sometime interacting with  D-Wave for some time, a relatively optimistic update in Nature.  Controversy remains.

D-Wave upgrade: How scientists are using the world’s most controversial quantum computer
Scepticism surrounds the ultimate potential of D-wave machines, but researchers are already finding uses for them. ... by Elizabeth Gibney

How does it work, how is it different?  In Nature:

" ... Unlike other quantum computers, D-Wave is suitable only for solving certain tasks, known as optimization problems. To find optimal solutions, researchers first put qubits, made of superconducting loops, into their lowest energy state, in which each is in a quantum superposition of both ‘on’ and ‘off’. Magnetic fields that represent the problem then gently nudge this state towards a new one — a process known as quantum annealing. The state evolves while maintaining its low energy such that when it eventually ‘collapses’, it should leave qubits in the best configuration for solving that problem. Because the system sifts every possible answer at once, in theory it could be a faster way to resolve problems that, when solved classically, get exponentially harder with each added variable. But posing research questions in a form that the machine can handle often means using several qubits to represent a single variable, limiting the size of the problems it can handle. ... "  

Digital Whiteboards

Over the hears have met with a number of companies aiming to make meetings more effective, and attended many meetings.  Have yet to see it well done. Usually the process used itself is not effective. Most often now, its taking pictures of quite standard boards.   Cisco tries again.

Cisco's Affordable Spark Board Wants To Change How You Conduct Meetings    This plug-and-play wireless screen works as a wireless presentation display, digital whiteboard, and video conferencing tool. .... "

Wednesday, January 25, 2017

Pulse Response Biometrics

Had never heard of this biometric, but as a student of 'biometrics' was intrigued.   No idea how unique such a signature is.  Article is non technical.

Authentication Using Pulse-Response Biometrics
By Ivan Martinovic, Kasper Rasmussen, Marc Roeschlin, Gene Tsudik 
Communications of the ACM, Vol. 60 No. 2, Pages 108-115 .... 

We propose a new biometric based on the human body's response to an electric square pulse signal, called pulse-response. We explore how this biometric can be used to enhance security in the context of two example applications: (1) an additional authentication mechanism in PIN entry systems, and (2) a means of continuous authentication on a secure terminal. The pulse-response biometric is effective because each human body exhibits a unique response to a signal pulse applied at the palm of one hand, and measured at the palm of the other. .... " 

Year of the Virtual Assistant


Why 2017 Will Be the Year of the Virtual Assistant
January 24, 2017,  by Danny Vena, Motley Fool

In its annual mobile apps survey, Research firm Gartner predicted that by 2019 virtual assistants will account for 20% of all smartphone interactions . With that level of user engagement,  predictably all of the largest technology companies are invested in the segment and each brings a different strength to the field.  .... " 

Sewing Robots

Most interesting direction,  change the materials to conform to the robotics.  I recall seeing some of the same videos which showed how human-complex sewing is.

It was pointed out to me that the precise alignment of the placement of the garment is not shown, and this is difficult, and still a challenge.

Videos at the link.    In FastCompany: 

Is This Sewing Robot The Future Of Fashion?
Startup Sewbo has figured out how to get a machine to sew an entire garment, and it may finally push clothing factories to fully automate. ....

 .... In the past, companies tried to create complicated mechanical devices to emulate the way a human sews, which "is a very difficult and complicated approach," explains Zornow. Instead, he took a different tack, manipulating the materials to make them compatible with robots.

He realized that if he stiffened the fabrics by drenching them in liquid polymers, they could be turned into thermoplastic composites and treated like hard materials. And robots need hard materials.

"They’re stiff as a board, but they can be molded: You can apply heat and reshape them, and when they cool down, they’ll hold their shape," explains Zornow. The machine sews through the stiffened fabric to produce a perfectly finished product. (The process can be used with any sewing machine and most robotic arms, which generally cost about $35,000.) Afterwards, the polymers can be easily washed off with water, no detergent necessary.  ... " 

On Google VR

Good Wired piece on how Google is looking at VR, where the user becomes the interface. The Inside Story of Google's Daydream, where VR feels at home  ... " 

Tuesday, January 24, 2017

CSI Talk: IT Technology Trends in 2017

From the CSI Group:

" .... This a reminder about our Cognitive Systems Institute Speaker Series on Thursday, January 26, 2016 at 10:30 am ET US (7:30 am PT).  Our speaker is William Chamberlin from IBM Market Development and Insights, who will be presenting "IT Technology Trends in 2017... and Beyond."

Please point your web browser to https://apps.na.collabserv.com/meetings/join?id=2894-8491   password=cognitive.   Use audio on computer or dial 855-233-7153 (other countries numberhere) PIN Code: 43179788    Non-IBMers please use the "guest" option instead of entering your email.   

Slides.   Talk.

Please find the schedule of presenters here for the next several calls.   A link to slides and a recording of each call should be available on the CSIG website (http://cognitive-science.info/community/weekly-update/).   We encourage those who join the calls to add questions and comments to the LinkedIn Discussion Group https://www.linkedin.com/groups/Cognitive-Systems-Institute-6729452  and please ask questions at the end of the call.

Thank you and please let me know if you would like to be a presenter on an upcoming call!  

Dianne Fodell

IBM Global University Programs
fodell@us.ibm.com   .... "  

Packaging and the Internet of Things

Packaging was always been an interest of our innovation group.  Elements of integrating packaging and the net have been talked for a long time.   It is starting to happen in interesting ways.  In Adage:

Packaging Deals Bring Internet of Things to Crackers, Hairspray and Tomato Paste   By Kate Kaye.
Headlines about connected products often hype the latest consumer electronics, seemingly out-of-reach home appliances, and the occasional internet-enabled bottle of whisky. In the past year, however, the near future of connected products became decidedly more mundane. Recent partnerships have begun the process of turning everyday items from hairspray to tomato paste into digitized, identifiable, data-generators that can be tracked from the factory to the pallet to the store to the consumer. ... " 

Voice Assistant Apps

Somehow don't think assistant apps are as important, never hear them mentioned when someone describes a good experience with assistants.  Still this is intriguing view of how they are used. 'services' in assistants are still not very good.   That is in part caused by an architecture that does not support intelligent interaction.   Still looking for that

Robotics in Technology Review:
AI Voice Assistant Apps are Proliferating, but People Don’t Use Them
Limited to just audio, users of Amazon’s Alexa and Google’s Assistant don’t seem to stick with apps that run on their helper’s software. Would a screen help?     by Jamie Condliffe   .... " 

See my list of assistants and tracking.

Mattel Aristotle by Nabi

Has some regulatory issues when interacting with children in the US:

" ... Mattel Aristotle is an Amazon Echo that understands your kids, too

It's a baby monitor. It's a voice-activated smart assistant. It answers your questions, and maybe your kids' questions, too. It orders more diapers when you run out and soothes babies back to sleep automatically. It plays with your kids. It could be the most exciting thing toymaker Mattel has ever produced.

It's called the Aristotle, and it's not just an Amazon Echo clone: According to interviews with Bloomberg, FastCo Design and USA Today, the device is a fully functioning Amazon Alexa assistant that can answer all the same adult questions and has all the same smart home capabilities -- but say "Aristotle" instead of "Alexa," and it can also summon a different voice assistant that can interact with your kids.   ... " 

See my list and links to this and other personal assistants.

Monday, January 23, 2017

Areas of AI to Watch

Very good piece by Nathan Benaich,  Covers at the right Tech savvy management-introductory level a number of AI oriented techniques.  I like these because each is more problem than technology facing.  Also brings in practical decision approaches we used in operations research.   Sections on each of these, have used a number of these, though sometimes under different names.   Note the less data article, critical.

Reinforcement Learning  
Generative Models
Networks with Memory
Learning from Less Data  and Building Small models  (Transfer Learning) 
Hardware for Training and Inference
Simulation Environments

Mentions of many companies and researchers involved in each area,  largely non-technical descriptions of each area.   Good read for all with strong interest in this area.

Nathan has " ...  an analytical digest of artificial intelligence and machine learning news from the technology industry, research lab and venture capital market  ...  ".     Link to it in the article,  I have subscribed.

Locating People in the Wild

Located beacons can detect people's mobile IDs in store and even on the street.  A powerful move to establish consumer journey.  I had heard of related methods some time ago, by testing and ID'ing people just walking into a store.

Programmatic Advertisers Get Even More Precise Consumer Location Data With New Pact   By  Kate Kay  In Adage.

Is Data the New Oil?

Interesting idea.  But we ingest in a different way than we do fuels.   While the chemistry of petroleum is well known and constant, the chemistry of required data is infinitely variable and constantly changing, so their general use is challenging,  which would lead to better understand data as an asset. Article and discussion in Linkedin.

AI is Changing the Retail Industry

How AI is Changing the Retail Industry for Consumers

Rick Delgado   

Artificial Intelligence is an up and coming technology system that is constantly enhancing different aspects of various industries. Some of these new areas where Artificial Intelligence has made a positive impact include traffic lights, healthcare, marketing strategies and commercial security solutions. Artificial Intelligence has recently found its way into the retail world. From conversing with consumers to developing improved management strategies, AI holds a promising future for creating efficient consumer experiences and transforming management systems for retail stores. ... "

Building Networks for Ecosystems

Intriguing piece about pipelines and platforms.  We discussed methods of building supply chains by evolving them.  Now with machine learning?

Building Networks for Scale: From Pipeline Models to Platforms, Ecosystems, and Machine Learning

Like Building Networks for Scale: From Pipeline Models to Platforms, Ecosystems, and Machine Learning Building Networks for Scale: From Pipeline Models to Platforms, Ecosystems, and Machine Learning
Eric L. Miley, MBA

here are many reasons why new ventures, established businesses, and large initiatives fail. They often fail to connect with investors, fail to reach the influencers that are needed to spread key messages, and ultimately fail to reach their stated goals or objectives. Most of us have seen promising ideas fizzle out as teams and organizations look to execute and implement ambitious visions. Of the efforts that do succeed, many tend to under deliver on their promises, waste too many critical resources, and take much too long to scale. In this article, I suggest that many of the problems that organizations face when seeking to grow and scale their most important work can be directly attributed to a failure to harness the power of their networks.

Traditional strategies often fail to use three key tools that are quickly becoming a necessity for any viable organization, change initiative, or large-scale project. These key tools are a robust platform, an engaged ecosystem, and the infusion of machine intelligence. In many ways, these are the three ingredients that organizations should be focusing on if they hope to effectively carry out their most important work. When understood, built, and managed properly, these tools will enable your teams to focus on the important tasks of innovating, problem solving, and delivering the value that is critical for the long-term success of your organization. .... " 

An AI on the Board

Is this even close to reasonable?  Depends how you define them ....   You could make the case that computers run many companies, but at a low level.   They provide information that leads to logic, that is exercised (mostly) to lead to decisions being made.  Course,  boards operate in very different ways ... and often cannot be said to actually 'run' the company the way a operations manager does.  Examples of apparent close coupling below are very interesting.

Could An AI Robot Run Your Company?    by Bernard Marr    

Robots — or more specifically, artificial intelligence — have become a major part of our modern, connected lives. We ask Siri for recommendations on where to eat, ask Google for directions to the restaurant, ask our car to play music or make a phone call, ask our house to turn on the lights and the heat before we get home.

But would you let an AI robot run your company?

One already does. Deep Knowledge Ventures, a Hong Kong-based life science venture capital company, has already appointed an AI robot to its board of directors. The robot, called Vital (Validating Investment Tool for Advancing Life Sciences) was designed to analyze trends in the data of certain companies in an effort to predict successful investments.

Vital is an equal member of the board and is expected to eventually be given an equal vote in all financial decisions made by the company.  .... " 

Sunday, January 22, 2017

Talking Tom Reaps Eyeballs

A stark reminder that the most desirable aspect of the Web is still getting eyeballs in the door, and then playing videos.  'Talking Tom'  seemingly trivial, with some interactive botlike aspects, has gotten some 10.3 billion video views.  How might we use these designs?

Talking Tom maker Outfit7 confirms it has been sold to United Luck Consortium for $1 billion   by  Dean TakaHashi in Venturebeat: ... " 

Display Cards for Alexa

One of the criticisms of voice-only assistants is there is no provision for visual display, and often visual is just better than reciting an answer over a speaker.  For example if its a recipe I want to see a picture of it.   Now announced, a feature in the developer forum for screen based products to provide screen cards.  Like the cards you get on your smartphone app when you are playing music.  Details   Via David Isbitski:

Saturday, January 21, 2017

Marketing Campaign

An intriguing example, with statistics,  that looks like a process model of how to do marketing campaigns.  Like to see this further built out with more detail, so it could be learned from in other domains.

In Think with Google:
" ... A Budweiser case study, told from three different viewpoints: brand, creative, and media. Hear the unique perspectives each team brought to the table to make Budweiser’s World Series campaign a home run.  ... " 

Friday, January 20, 2017

Tips for VR Content

In Think with Google: 
 " ... It's one thing to make your VR content accessible. But how can you ensure that your audience engages? Our recent research identified three key insights for "sticky" VR content that you can carry into your highly non-virtual media plan ... 

We recently shared some tips for how to make VR content accessible to more users, based on takeaways from a YouTube Red Originals campaign. One of the biggest upshots was that VR content doesn't need to have sky-high production values to be immersive and compelling. That got us thinking about other traits of winning VR content. .... " 

AI Building AI

It has been long suggested that this would occur.  This is still relatively hard, but we have seen apps like Watson Analytics show how it could be done.  Though that is about data science, and this further embodies the logic and process of AI.  In Technology Review:

 AI Software Learns to Make AI Software
Google and others think software that learns to learn could take over some work done by AI experts.
by Tom Simonite .... "

Artificial Fingertip Feels

Artificial fingertip that ‘feels’ wins Harvard's robotics competition

TacTip, a 3D-printed tactile sensor  (University of Bristol) 

An open-source 3D-printed fingertip that can ‘feel’ in a similar way to the human sense of touch has won Harvard University's international Soft Robotics competition for its contribution to soft robotics research.  ....   "

Thursday, January 19, 2017

Scan and Go Again

Scan and Go Again  In Retailwire.

by Matthew Stern with expert comments.

Some ideas were ahead of their time and that’s how Walmart appears to see its Scan & Go mobile app. The retailer is bringing the discontinued mobile checkout app back into the pilot phase.

Walmart began a pilot run of the reintroduced app in a single store in Rogers, Arkansas, according to Tech Times. The app, released to the Google Play store on January 9, enables shoppers to scan the barcodes of items with their smartphones as they put them in their baskets. They then check out via mobile shopping cart and need only present their receipt to leave the store. The app also keeps a running tally of the cost of groceries, so shoppers can track their spending throughout their shopping trips.

The first run of Walmart’s Scan & Go app began in September of 2012 and was eventually expanded to a few hundred stores. Walmart shut down the pilot in 2014 because customers in the test locations could not figure out how to use it.  .... " 

Help Every Child Learn to Read

Curious Learning.

Help Every Child Learn to Read
2 Years Of Early Literacy Education, Delivered By Smartphone Or Tablet.


Created At MIT, Backed By Research. .... 

Revealing the Limits of AI

So what can AI really do today?  Here a list of questions which aim to indicate the current limitations. Includes inputs by Roger Schank,  AI researcher we have followed for some time. Worth understanding if you are interested in implementing something.  Does not mean we can never do these, but that they are still hard to implement.   In CDN: 

Questions that Reveal the Current Limits of AI

Using Mobile Platforms for Early Literacy Education

This was quite good,  included mention of the use of screen robots to provide learning companions to children.   Quite a good example of the use of smartphones to teach up to Pre-K and K levels.   Uses some AI avatar capabilities.     Already has worked with 100K children.  Includes both cognitive App development and App selection for specific learning.

." ... Our speaker today was  Taylor Thompson from Curious Learning, a MIT spin-off, who will be presenting "Curious Learning: Using a Mobile Platform for Early Literacy Education as a Research Sandbox."  ...

 ....   A link to slides and a recording of each call should be available on the CSIG website (http://cognitive-science.info/community/weekly-update/).   We encourage those who join the calls to add questions and comments to the LinkedIn Discussion Group https://www.linkedin.com/groups/Cognitive-Systems-Institute-6729452 and please ask questions at the end of the call.

Thank you and please let me know if you would like to be a presenter on an upcoming call!  

Dianne Fodell

IBM Global University Programs

AI Resolving Customer Disputes

Note the use of personality profiles.  How accurately can this be done?

Can AI resolve customer service disputes?  by Matthew Stern  

A phone call with a customer service (CS) representative can devolve into a nightmare for any number of reasons. But to the extent that problems are caused by a particular customer and CS representative just not getting along, one startup aims to ease the tension with artificial intelligence (AI).

The startup, called Afiniti International Holdings, has created an AI-based system that builds a profile of a customer service caller based on aggregated information. When routing a call, it uses the profile to determine which customer service associate will most likely achieve a successful interaction, according to Engadget. ... " 

Wednesday, January 18, 2017

China Doubling Down on AI

Have been in particular following their assistant and chatbot work, impressive so far.    Further indications that work is ongoing in augmented reality.  In Technology Review:

In 2017, China Is Doubling Down on AI
The country’s Internet giants are focusing on AI research, and domestic venture capital funding is pouring into the field.   by Jamie Condliffe    

This year, China looks set to make larger waves than ever in artificial intelligence and augmented reality.

The nation’s search giant, Baidu, is leading the charge. Already making serious headway in AI research, it has now announced that former Microsoft executive Qi Lu will take over as its chief operating officer. Lu ran the applications and services division at Microsoft, but, according to the Verge, a large part of his remit was developing strategies for artificial intelligence and chatbots. In a statement, Baidu cites hiring Lu as part of its plan to become a “global leader in AI.”

Meanwhile, Baidu’s chief scientist, Andrew Ng, has announced that the company is opening a new augmented reality lab in Beijing. Baidu has already made progress in AR, using  computer vision and deep learning to add an extra layer to the real world for millions of people. But the new plans aim to use a 55-strong lab to increase revenue by building AR marketing tools—though it’s thought that the company will also consider health-care and education applications in the future. .... " 

Wharton Overview of Blockchain

Good, non technical overview of blockchain in Knowledge@Wharton.  Very nontechnical piece that is worth reading.  Aims to address the question: " ...   Is Blockchain the Next Great Hope — or Hype? .... ".   Largely lands on the positive side, with some discussion of the risks involved.    Discussion is good too.  Worth passing on as a non technical introduction.

Connect + Develop Looks for Innovation.

Have not mentioned it very often, but Connect + Develop is P&G's Open innovation capability.  I get its flyers presenting specific innovation opportunities.  See the latest flyer here.

" ... Below is another Connect + Develop opportunity from Procter & Gamble that may be of interest to you.

As background, P&G's Connect + Develop group seeks to partner with the best innovators everywhere, and employs a number of approaches to connect with a network of top innovators around the world. 

More than half of new product initiatives at Procter & Gamble involve significant collaboration with those outside P&G walls. They partner with small companies, multi-nationals, individual inventors, and in some cases, even competitors, to bring game-changing innovations to market. Connect + Develop is at the heart of how P&G innovates.   ... " 

See their full site here.

Inventory Sensing Refrigerators

A topic we took up in innovation.  Below repeated in the Verge.     If you could determine inventory and predict usage.   You could just reorder goods.  Of course the frig is not the only storage place in the kitchen.    So some suggested that you could RFID scan things as you discarded/recycled their packaging.  Several such 'smart' trashcans were marketed.   Samsung has a frig now, apparently for sale,  definitely very expensive, that says it can solve the inventory problem.    Given Amazon's recent efforts, like the Dash Button, would think that they would be interested in the approach.  Does the consumer want this?   Or does that really matter?   In The Verge: 

Wanted: An Amazon fridge that automatically reorders food
 January 18th, 2017  by Thomas Ricker .... 

See also, The FridgeCam.

Reviewing Turings Enigma

Reviewed the book:  Alan Turing: The Enigma  by Alan Hodges. Hodges also has a site that provides resource materials and updates to the book at   http://www.turing.org.uk.     See the movie, The Imitation Game, now available on Netfllix, generally based on the book.

Tuesday, January 17, 2017

IBM Announces Cognitive Principles

Announcement by IBM in Davos.

January 17, 2017 | Written by: IBM THINK Blog
Categorized: Cognitive Computing

We are in the early days of a promising new technology, and of the new era to which it is giving birth. This technology is as radically different from the programmable systems that have been produced by the IT industry for half a century as those systems were from the tabulators that preceded them.

Commonly referred to as artificial intelligence, this new generation of technology and the cognitive systems it helps power will soon touch every facet of work and life – with the potential to radically transform them for the better. This is because these systems can ingest and understand all forms of data, which is being produced at an unprecedented rate. Cognitive systems like IBM’s Watson can reason over this data, forming hypotheses and judgments. Most importantly, these systems are not simply programmed, they learn – from their own experiences, their interactions with humans and the outcomes of their judgments.

As with every prior world-changing technology, this technology carries major implications. Many of the questions it raises are unanswerable today and will require time, research and open discussion to answer. But at IBM, we have learned something over 100 years of inventing and introducing transformative technologies and of guiding their responsible adoption and use. This experience has taught us that it is both pragmatic and wise to establish principles to guide what we develop and bring to the world, and how we do so.   ....  "   

List of principles at the link .... 

3D Data Optimization by Simplygon

Interesting play,  taking a close look.    Initially used for game applications, and I expect further used for dealing with multiple dimensional data in interactive 3D displays and virtual reality.  Design applications are likely early applicatinos.   Good examples at the Simplygon site.  In ZdNet: 

Microsoft acquires 3D data-optimization vendor Simplygon
Microsoft's latest acquisition, 3D data-optimization vendor Simplygon, is part of Redmond's strategy to make 3D part of future users' experiences.    By Mary Jo Foley

" .... Simplygon will help Microsoft simplify the process of capturing, creating, and sharing information in 3D, Tsunoda said. The Simplygon technology will complement the new Paint 3D application (codenamed "Beihai") and new online creator community, Remix3D.com.

Simplygon was developed by Donya Labs AB in Sweden, a company developing "automatic 3D data-optimization solutions." .... ' 

IEEE on the Kuri Robot

Previously mentioned.  The Kuri Robot, by Bosch and Mayfield,  scheduled to be out later this year.  ... Mayfield Robotics Announces Kuri, a $700 Home Robot   By Evan Ackerman.   Appears to be a solution that attempts to create emotional interaction with the household.   That part remains to be seen.  The concrete information supplied makes it appear more like a mobile security camera.

Predicting Machine Problems with Deep Learning

Maintenance prediction is something we worked on.  Here a related application that has many possibilities.   How about very rare breakdown occurences?  In IEEE Spectrum: 

Deep Learning AI Listens to Machines For Signs of Trouble    By Jeremy Hsu

" .... The service of 3DSignals, a startup based in Kefar Sava, Israel, relies on the artificial intelligence technique known as deep learning to understand the noise patterns of troubled machines and predict problems in advance. 3DSignals has already begun talking with leading European automakers about possibly using the deep learning service to detect possible trouble both in auto factory machinery and in the cars themselves. The startup has even chatted with companies about using their service to automatically detect problems in future taxi fleets of driverless cars. .... " 

Leadership in a Cognitive era

Discussion starting in the Cognitive Systems Linkedin Group.   Join us in the conversation.

Points to a Forbes article.
 How will leadership change in the Cognitive era?   Companies like IBM, Facebook, Microsoft and Amazon are bringing cognitive computing to the masses ....     By Jacek Frankowski

I comment:  Some thought provoking ideas in the article. But insufficiently covered. Will this lead to more following and less leading? Being led by intelligent algorithms? I don't think so. It will mean that we need to better understand the implications of more data driven and quicker decisions. Cognitive (Voice, reasoning, collaboration, assistance .... ) means that we will have the opportunity for closer connections to what is happening. Need to string it all together by better understanding the process. ..... " 

Raspberry Pi Upgrade

Raspberry Pi upgrades Compute Module with 10 times the CPU performance ... Compute Module for embedded computing gets first big upgrade since 2014.   by Jon Brodkin

 Am a big fan of this, its cheap.  Easy to get started. Would suggest that any child with a hint of interest in computing, electronics or the Internet of Things should get access to one of these.  Admit its a little obscure, and many will need some initial hand holding, but there is no better way to get a start.  Lots of resources on the internet and opportunities to collaborate.  In my era, we had breadboard systems where you could easily set up circuits and test them.  That gave me a lifelong appreciation of the details.   This is a similar way to gen up interest in useful skills.  See more in my tags below on Raspberry Pi.

Forbes on Uses of AI today

In Forbes:  Mostly well known, and can be said that most of these are very focused applications.  The assistant applications like Siri and Alexa, can be said to be appear general, but they are still not providing general intelligence, but more efficient usage of better sources of plentiful data and some kinds of service specific skills, like natural language understanding and pattern recognition.

Monday, January 16, 2017

Trends in Digital Transformation

Four Trends in Digital Transformation

In  CustomerThink by Jacob Morgan

If one of the main goals of the future of work is to increase productivity and collaboration, there’s a great place to start: digital transformation. Many companies are embarking on a digital transformation in an effort to connect employees and customers around the world digitally. The future of work and digital transformation are both vital to each other’s success, and they can work together to help organizations prepare for a new wave in the workforce. A digital transformation can provide employees the tools they need to create the best work environment for the future.

Digital transformation builds on the current state of global connectivity and can include a wide variety of features. However, according to Adam Warby, CEO of global technology solutions firm Avanade, there are four main trends being seen in digital transformations across industries:  .... " 

Course about AI in the Enterprise

Ajit Jaokar of Futuretext has an upcoming enterprise AI course.
I like the idea of aiming this at the enterprise, which is rarely done,  but needs to be done more.   Its not just about solving the problems, it is about making it easier to do in the future.  The details in the link below, still some room as I understand it.  He writes:

Hello all

We had a great response to the Enterprise AI course and about to close it now
If you are interested in signing up, please let me know

Course outline 


kind rgds

Simulation Against Real World Data for Learning

Points to how games can act as simulations of the world, and emulate their operation against real world data.    Experimented with this in the supply chain space.  In TechnologyReview:

Robot Cars Can Learn to Drive without Leaving the Garage
Playing video games and surfing Google Street View can teach software a lot about driving. by Will Knight  .... 

" ... Researchers at Princeton University recently developed a computer vision and mapping system that gathered useful information about the physical properties of roads by studying Google Street View and comparing the scenes to the information provided in open-source mapping data. This allowed it to, for example, learn where the edges of an intersection should be based on images captured by Google’s mapping cars. .... " 

Sunday, January 15, 2017

Service Science Readings

Service Science Innovation:  Wanted to remind people of this resource, it has recently been updated.   I have contributed with some of the work we did in enterprise.  Feel free to dive in and contribute if you like.  This is particularly relevant to the use of AI  and expertise based systems to augment service systems.  And one common kind of such a system has an assistant model.   Location and recent updates.    @The_ISSIP  #ISSIP   Via Jim Spohrer.

Complex Systems Similarity through Fractal Folding

By folding fractals into 3-D objects, a mathematical duo hopes to gain new insight into simple equations. .... " 

The only direct application I ever saw of fractals in industry was for an unusual packaging application.    This article suggests that they can be used to suggest some similarity between otherwise different equations.  And maybe then used to solve them in new, perhaps faster ways?   But no examples are given, so am hard pressed to see it.  But good luck, and it is always astounding to see beauty arise from such simple statements of mathematics.  Why?  - FAD

There are lots of Apps and packages online for experimenting with fractals, if you have a child that is even a little math inclined, easy to demonstrate the ideas. -   FAD 

Benefits of RFID Data in Retail

Recently received, quite interesting. 75 pages based on Macy's data.
From the Platt Research Institute.

The new PRI Working Paper, Quantifiable Benefits and Analytical Applications of RFID Data, analyzes historical data provided by Macy's related to its RFID program. RFID technology provides unprecedented visibility into the location of retail merchandise and, therefore, provides benefits that may include reduced inventory requirements, enhanced omni-channel fulfillment, influence sales, improved customer satisfaction, reduced markdowns and labor costs, as well as improved supply chain coordination. ... " 

The research detailed in this 75-page Working Paper is focused on four Use Cases:

Display Audit.
Inventory Accuracy.
Single Unit Fulfillment.
Back to Front.

The use of RFID technology in Macy's Stores has resulted in improved display compliance, inventory accuracy, and making single merchandise units visible and available for sale.

"To our knowledge, this is the most extensive set of data made available on the quantifiable attributes associated with RFID within a retail store," said Steven Keith Platt, Director and Research Fellow, Platt Retail Institute, and Research Director, Retail Analytics Council, Northwestern University. "We would like to thank Macy's for its ground-breaking work with item level RFID and its participation in this study."   .....' 

Humans are Still at Work

In the HBR.    For now the humans are clearly still in the loop. we can learn from some areas, like autopilots or process control here.    Helping them effectively be in the loop is part of our near term task.    It means integrating a conversation, leading from problem to solution, that may include a number of machines and humans.  We do this already, when an accountant uses a spreadsheet.  Now how will each human or machine most efficiently and credibly contribute in a problem solving process?

The Humans Working Behind the AI Curtain
Mary L. Gray,  Siddharth Suri

There is a human factor at work in tasks promoted as artificial intelligence (AI)-driven, in the form of people paid to respond to queries and requests sent to them via application programming interfaces of crowdwork systems, write Microsoft Research scientists Mary L. Gray and Siddharth Suri. "The creation of human tasks in the wake of technological advancement has been a part of automation's history since the invention of the machine lathe," they note. "We call this ever-moving frontier of AI's development the paradox of automation's last mile: as AI makes progress, it also results in the rapid creation and destruction of temporary labor markets for new types of humans-in-the-loop tasks." 

Gray and Suri predict the enhancement of human services by AI will augment daily productivity, but present new social challenges. "The AI of today can't function without humans in the loop, whether it's delivering the news or a complicated pizza order," the researchers note. Technology and media companies therefore employ people to perform content moderation and curation, while many jobs are outsourced overseas and paid a low, flat rate. "This workforce deserves training, support, and compensation for being at-the-ready and willing to do an important job that many might find tedious or too demanding," according to Gray and Suri.  .... " 

Another Face and Voice for the Alexa Engine

A new kind of speaker, Jam Voice,  it's based on the Alexa engine, but it won't listen.    Another cheaper solution in the space.   How different than that, don't know.  In theVerge.

Saturday, January 14, 2017

EU to Give Robots Personhood?

Considerably premature.

Give Robots 'Personhood' Status, EU Committee Argues   
by The Guardian  in the CACM.

The European parliament has urged the drafting of a set of regulations to govern the use and creation of robots and artificial intelligence, including a form of "electronic personhood" to ensure rights and responsibilities for the most capable AI.  .... " 

McKinsey on Automation

The broader issues of how we automate is important.  As methods like AI start to be inserted into the business process, clear choices will have to be explored, made and continually monitored and improved.  A good article in McKinsey: 

" ... Automation is happening, and it will bring substantial benefits to businesses and economies worldwide, but it won’t arrive overnight. A new McKinsey Global Institute report finds realizing automation’s full potential requires people and technology to work hand in hand. .... " 

Forty Techniques for Data Science Analytics

An excellent list and pointers to introductory articles on over forty data science techniques.  Have mentioned these before, worth the repeat.  By Vincent Granville.  The practitioner should know the basics of many of these.  All very well done.  From DSC.   Nicely done.   Join DSC and get their newsletters,  which are useful.

Friday, January 13, 2017

Alexa Becoming the Voice of the IOT?

 Maybe it won't be the 'brain', but will it be the voice of the IOT,  and through the IOT to many kinds of goods and services?   I would still bet on Google or others sculpting the brain,  but Amazon is leading the pack on voice services and existing infrastructure.

" ... Amazon has sold “millions” of its Echo Bluetooth speakers with Alexa inside. Now that hundreds more Alexa-enabled devices are coming to market, the benefits to Amazon will grow, and quickly.

Amazon smartly made Alexa an easy to work with, flexible program, much like Apple did with the iPhone to spur app development. It is essentially creating an operating system for the home, one that will connect the myriad of smart devices in the Internet of Things to each other and to Amazon. ... " 

Self Organizing Robotics

Had not realized that Harvard was working on this.  Potential medical applications at least.   In the CACM:   " ... Harvard University's Self-Organizing Systems Research group has developed a "large-scale robot collective" that can self-assemble into different shapes. ....  The system is based on a subtractive approach, instead of the additive approaches to autonomous self-assembly robotics. ... " 

Electric Cars Going Mainstream

A question many of us have been thinking about.    In Knowledge@Wharton, a good economic oriented view of the direction.    Innovation: When Will Electric Cars Go Mainstream?  
Costs, Range and Infrastructure Still Impose Limits ... " 

January-February 2017 Analytics Magazine

Informs January-February 2017 Analytics Magazine
An issue with emphasis on the IoT and its evolution.  Such as on related business opportunities.

Sensors in the Internet of Things

Good early introductory and pictorial piece on the interaction of sensors and sensing for the Internet of things.  Pointers to more resources.

Thursday, January 12, 2017

Pinterist uses Deep Learning

Nice example of a use of deep learning.

 Pinterest uses deep learning AI for better pin recommendations 
By Eric David

Pinterest Inc. revealed today that its Related Pins feature will now be powered by deep learning neural networks, making it even easier for users to find all of the dessert recipes and pictures of things in mason jars they could possibly want ... "

Autonomous Collaborating Micro-Drone Swarms

The autonomy of such drones and drone swarms indicates an unprecedented depth of decision making driven by AI.  Implications are very broad.

Pentagon successfully tests micro-drone swarm  In mPhys: 
The Pentagon may soon be unleashing a 21st-century version of locusts on its adversaries after officials on Monday said it had successfully tested a swarm of 103 micro-drones. The important step in the development of new autonomous weapon systems was made possible by improvements in artificial intelligence, holding open the possibility that groups of small robots could act together under human direction.

Military strategists have high hopes for such drone swarms that would be cheap to produce and able to overwhelm opponents' defenses with their great numbers.

The test of the world's largest micro-drone swarm in California in October included 103 Perdix micro-drones measuring around six inches (16 centimeters) launched from three F/A-18 Super Hornet fighter jets, the Pentagon said in a statement.

"The micro-drones demonstrated advanced swarm behaviors such as collective decision-making, adaptive formation flying and self-healing," it said. ...." 

And in Ars Technica.

Jill Watson at Ga Tech

Have worked with many TA's myself, so this is interesting.  It also relates to chatbots too, lessons to are to be learned here.   As in all tests of this type I wonder about the difference between providing value to someone versus fooling them.  Not the same thing. The goal can sculpt the results.    Read the full piece.

Jill Watson, Round Three.
The Georgia Institute of Technology (Georgia Tech) is beginning its third semester using a virtual teaching assistant (TA) system, called Jill Watson, in an online course about artificial intelligence (AI).

Jill, which is implemented on IBM's Watson platform, was first used last spring to successfully answer particular types of frequently asked questions without the help of humans.

Georgia Tech professor Ashok Goel told the students at the beginning of the semester some of their TAs may or may not be computers. "Then I watched the chat rooms for months as they tried to differentiate between human and artificial intelligence," Goel says.   ... " 

Full list of advisory systems being followed.

ACM Statement on Algorithmic Transparency and Accountability

The well known ACM professional society issued a statement about algorithmic transparency and accountability today.  I have been a member and participant for many years.   Algorithms have been part of computing forever, but only recently have they been closely examined regarding their implications, especially as they interact with the public.  And we interact with them every day.

 Algorithms are models, and inherently have bias.  They make and position decisions for and with us. This will further expand with the use of AI.   The ACM statement, the first few paragraphs below, the complete document is at the link, does an excellent job of laying out the problem, and their professional position.  Well stated:

Statement on Algorithmic Transparency and Accountability
Computer algorithms are widely employed throughout our economy and society to make decisions that have far-reaching impacts, including their applications for education, access to credit, healthcare, and employment.    The ubiquity of algorithms in our everyday lives is an important reason to focus on addressing challenges associated with the design and technical aspects of algorithms and preventing bias from the onset.

An algorithm is a self-contained step-by-step set of operations that computers and other 'smart' devices carry out to perform calculation, data processing, and automated reasoning tasks. 

Increasingly, algorithms implement institutional decision-making based on analytics, which involves the discovery, interpretation, and communication of meaningful patterns in data. Especially valuable in areas rich with recorded information, analytics relies on the simultaneous application of statistics, computer programming, and operations research to quantify performance.

There is also growing evidence that some algorithms and analytics can be opaque, making it impossible to determine when their outputs may be biased or erroneous. Computational models can be distorted as a result of biases contained   ..... " 

Chatbot Retention Problem

The open question is, how well will people accept non human online interaction?    Bots can be novel, but that is not enough.  Bottom line, they need to provide a faster and more accurate interaction than talking to a human or reading a list of solutions.  They also need to be able to utilize the context of a problem as much as possible. I want a system to detect a problem and solve it without interaction if at all possible.  Bring on intelligent autonomy.   Below some good general guidelines and bot examples.

3 tips on improving chatbot retention   by Stefan Kojouharov  in Venturebeat: 

Chatbot retention has been a real problem. It’s so poor that most people don’t even get past the first two messages. According to İlker Köksal, the CEO of BotAnalytics, the initial drop-off is huge: “About 40 percent of users never get past the first text, and another 25 percent drop off after the second message. Daily retention rate is at a paltry 1–2 percent, and the monthly retention rate for bots isn’t much better, sitting at about 7 percent.” Fortunately, after hacking for the better part of 6 months, a few bots — such as the weather bot Poncho — have found the light and are seeing awesome retention and engagement rates.

There is such a wide variety of chatbot use cases that it does not make sense to compare against the average. You might have a use case that solves a one-time problem, in which event you hope the user never has to come back (such as the DoNotPay lawyer bot). The best way to benchmark is by comparing your bot to mobile apps in your category. At the bare minimum, your goals should be to surpass them.  ... " 

Industrial Lab for IoT

National Instruments industrial IoT lab unites rivals

Cisco, HPE, Intel and others help sponsor the lab to show enterprises that IoT systems can work    By Stephen Lawson  

There are many companies vying to build the industrial internet of things, but the systems involved are so complex that those vendors also need to cooperate. A new lab at National Instruments, in Austin, Texas, is bringing some competitors together.

The NI Industrial IoT Lab opened on Wednesday and will house testbeds for applications like predictive maintenance, time-synchronized industrial networking and “microgrids” for renewable energy. It will also be a place where companies can show off joint solutions to customers. ... " 

Wednesday, January 11, 2017

Digimarc at the NRF

I had mentioned Digimarc and their advanced product marking technologies, recently in a post.   They will be present at the upcoming 2017 NRF show on January 15-17 in NYC.   Here is an invite and more information about their company.  See the tag below for more up to date information about them in this blog.

Intro to Neural Networks and Meta-Frameworks for Deep Learning with TensorFlow

This starts with the basics of neural networks, which I find very useful.  We worked with Neural networks for learning as early as the 80s.   Good straightforward overview of the topic.   Also addresses the kinds of learning and memory involved.  This will be crucial in the longer term to understand how we converse with AI enabled assistants.

Introduction to Neural Networks and Meta-Frameworks for Deep Learning with TensorFlow
Posted by Roger Strukhoff, Director of Research, in Machine Learning
Tags: Machine Learning, TensorFlow
With sample code and demos, this blog post highlights major topics covered at a recent TensorFlow webinar: what it takes to train a recurrent / convolutional neural network, four unique object types, meta-frameworks, etc.  .... " 

Why: A Guide to Finding and Using Causes

I brought this book up some time ago.    And as I read it am struck by how essential is is to understand cause and effect when working with data, analytics, machine learning and assistant services.  This is MORE important than understanding code or data structure.

The book does an excellent job, starting non-technically, and extending to explain recent research. It is worth at least a thoughtful scan.  If you don't understand this, your modeling can be dangerously invalid.   The link below goes to a free 23 page sample pdf of the book. After that, buy it for reference.

Why: A Guide to Finding and Using Causes
by Samantha Kleinberg

"Kleinberg expertly guides readers on a tour of the key concepts and methods for identifying causal relationships, with a clear and practical approach that makes Why unlike any other book on the subject. Accessible yet comprehensive, Why is essential reading for scientific novices, seasoned experts, and anyone else looking to learn more from data."  .... '

Samantha Kleinberg's Site.  and book ordering information.

Developing Services for the Google Home

Want to know what it takes to develop new services for systems like the assistants Amazon Echo and Google Home? Came upon this largely non technical and simplified set of posts that address the approach for the just released Google Home that I am testing for assistant use.   Worth a look to start to understand the process.  Starting to review myself.   I will report back as I can.

Transfer Learning for AI Projects

Had always thought that intelligence was about learning, so this concept struck me.   Note mention of improbable events and model correctness maintenance,  always of concern in such studies.  Technical.

'Transfer learning' jump-starts new AI projects
Machine learning, once implemented, tends to be specific to the data and requirements of the task at hand. Transfer learning is the act of abstracting and reusing those smarts

'Transfer Learning' Jump-Starts New AI Projects  in InfoWorld by James Kobielus

Abstracting and reusing knowledge gleaned from a machine-learning application in other, newer apps--or "transfer learning"--is supplementing other learning methods that constitute the backbone of most data science practices. Among the technique's practical uses is productivity acceleration modeling, which is viable when prior work can be reused without extensive revision in order to speed up time to insight. Another transfer-learning application involves the method helping scientists produce machine-learning models that exploit relevant training data from prior modeling projects.

This technique is particularly appropriate for addressing projects in which prior training data can easily become obsolete, which is a problem that frequently occurs in dynamic problem domains. A third area of data science in which transfer learning could yield benefits is risk mitigation. In this situation, transfer learning can help scientists leverage subsets of training data and feature models from related domains when the underlying conditions of the modeled phenomenon have radically changed. 

This can help researchers ameliorate the risk of machine-learning-driven predictions in any problem domain vulnerable to extremely improbable events. Transfer learning also is critical to data scientists' efforts to create "master learning algorithms" that automatically obtain and apply fresh contextual knowledge via deep neural networks and other forms of artificial intelligence. ... " 

Uncertainty in US Economy

In Knowledge@Wharton,  on uncertainty in the US economy

" ...  “The biggest risk we face is uncertainty. If you ask every business leader, their biggest concern is: ‘Whatever changes occur, just do them gradually. Let us adapt.’ They have been in a world of change for a long time. That’s not going to go away; we can’t take away change…. Their biggest concern is if something hits them, and they simply can’t or don’t have the time and the resources to adjust.”

Harker and Wharton finance professor Jeremy Siegel discussed the outlook for the U.S. economy in 2017 on the “Behind the Markets” show on Wharton Business Radio on SiriusXM channel 111. Siegel hosts the show with Jeremy Schwartz, director of research at WisdomTree. (Listen to the podcast at the top of this page.) ... " 

Tuesday, January 10, 2017

Pringles are Getting Louder

Kellogg bought Pringles canned chips from P&G in 2012 for $2.7 Billion.  Was slightly involved in the brand.   The new brand is pushing new and spicy tastes, and is selling well.  Strikes me that spicier foods continue to grow.  What is causing that?   See Pringles tag below for more.

Chatbots in Retail

In Retailwire.  Is 2017 the year of the retail chatbot?   Some good discussion.  We are not there yet,  there are still some issues to be worked out, but as long as they are focused on key issues and interactions, I think we are close.  " ... Yet many industry experts believe chatbots could become even more closely integrated into the shopper journey both online and in-store, particularly as consumers become more accustomed to the technology’s possibilities. ... ".    Further, read Peter Fader's cautious thoughts, he has a good understanding of potential analytic approaches.

GE Geneva: Understands 55 Billion Utterances

Examining how devices are attached to voice interfaces. Like to see this example of assistant systems, now how universal will such approaches be?

" ... Use Your Voice:  
Introducing voice-activated control of your GE WiFi Connect appliances with Amazon Echo

Meet Geneva:  Geneva is an Alexa skill that works with your Amazon Echo to take GE’s suite of WiFi Connect Appliances to the next level

Amazon Alexa
Geneva is available to download as a skill on the Alexa app. ... "