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

Monday, July 16, 2018

Model for Large-Scale 3D Facial Recognition

Can information in other dimensions include other useful business components of an image?

New Model for Large-Scale 3D Facial Recognition
By University of Western Australia 

University of Western Australia researchers have developed a three-dimensional (3D) facial recognition system based on the analysis of 3.1-million 3D scans of more than 100,000 people.

The team trained its FR3DNet system to learn the identities of a large dataset of "known" persons and then match a test face to one of those identities.

The research demonstrates that recognition performance on 3D scans is getting “better and more robust,” says the University of Western Australia's Syed Zulqarnain Gilani.

FR3DNet can identify faces in any pose, wearing glasses or a face mask, laughing or smiling, says Gilani, adding, “We hope that this research will help improve security on devices that use facial recognition to grant access to networks and systems.” ... " 

From University of Western Australia 

Byron Reese: The Fourth Age, Smart Robots, Conscious Computers and the Future of Humanity

Currently reading, especially interesting regards the influence of smart systems on jobs and work.  What will be out new role in this new age be?   Should we embrace or feat the age?  Nicely done

 The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity   by Byron Reese  They write: 

Our world up to recent times has been a Third Age world. While incredible innovation has occurred along the way, such as the development of steam and electric power and the invention of movable type, these were not fundamental changes in the nature of being human the way language, agriculture and writing were. With the exceptions of computers and robots, the innovations that we have observed have been evolutionary more than revolutionary. This is not to diminish them in the least. Printing changed the world profoundly, but it was simply a cheaper way to do something that we already could do. Detailed schematics of a biplane would have made sense to Da Vinci. But computers and robots are different. If we use them to outsource thought and motion, the very essence we are, then that is a real change, a Fourth Age.

“Reese frames the deepest questions of our time in clear language that invites the reader to make their own choices. Using 100,000 years of human history as his guide, he explores the issues around artificial general intelligence, robots,  consciousness, automation, the end of work, abundance, and immortality. As he does so, Reese reveals himself to be an optimist and urges us to use technology to build a better world.”  — Bob Metcalfe, UT Austin Professor of Innovation, Ethernet inventor, 3Com founder

“In The Fourth Age, Byron Reese offers the reader something much more valuable than what to think about Artificial Intelligence and robotics—he focuses on  HOW to think about these technologies, and the ways in which they will change the world forever. If you only read just one book about the AI revolution, make it this one.”  — John Mackey, co-founder and CEO, Whole Foods Market  .... " 

Microsoft Open Data

Brought to my attention:  Microsoft Open Data.  Their blog post about it.

A collection of free datasets from Microsoft Research to advance state-of-the-art research in areas such as natural language processing, computer vision, and domain specific sciences. Download or copy directly to a cloud-based Data Science Virtual Machine for a seamless development experience. ... "

Potentially very useful.  https://msropendata.com/

Concept and History of Algorithms

In my executive interactions since the beginning,  the definition of algorithm is often brought up.    Somehow it always seems mysterious.  Her a look at the definition and history of the idea. Its just a description of the statement of 'how do you do that?'   Which has been around for a long time.   Ever since someone has asked "How do you do that?"   With the implication its a good way, but that still depends on the author of the process.  Perhaps we chose a too complicated name.

Algorithms Have Been Around for 4,000 Years  By Herbert Bruderer 
A basic concept of computer science is the algorithm. An algorithm can be described as:

instructions for solving a task; a method for solving a problem; calculation rule, or, more precisely,
a finite sequence of generally (valid), unique, executable instructions (steps).
The technical term is named after the Persian mathematician Muhammad Ibn Musa al-Khwarizmi, author of a work on calculation rules (who lived around 780 to 850 AD). Examples from everyday life are recipes, handicraft instructions, rules of the game, instructions for use, score, pattern. .... "

Impact of Digital Life

Have seen several stark examples of this of late. And the integration of smart things will make this more important yet.

Stories From Experts About the Impact of Digital Life
By Pew Research Center

Over the years of canvassings by Pew Research Center and Elon University's Imagining the Internet Center, many technology experts and scholars have been anxious about the way people's online activities can undermine truth, foment distrust, jeopardize individuals' well-being when it comes to physical and emotional health, enable trolls to weaken democracy and community, compromise human agency as algorithms become embedded in more activities, kill privacy, make institutions less secure, open up larger social divisions as digital divides widen, and wipe out untold numbers of decent-paying jobs. ... " 

Amazon is Convenience

Convenience and consistency of results.  And over a relatively short period of time, re-setting expectations of delivery and immediate gratification.

How Amazon Delivers on Its Core Product: Convenience
Wharton's Katja Seim discusses her research on Amazon's fulfillment center network.


Amazon sells more goods than any one person could count – but the e-commerce giant’s true “core product” is convenience, and how quickly it can get an order from customers’ virtual shopping carts to their real-life doorsteps.

Part of what makes it so easy for Amazon to offer two-day or even same-day shipping to customers is its vast network of distribution centers, which are located across the U.S. and store and ship products to their final destinations. New research from Wharton business economics and public policy professor Katja Seim takes a closer look at how significantly expanding that distribution center network over the past decade has been key to Amazon’s growth strategy.

Seim recently spoke to Knowledge@Wharton about her paper, “Economies of Density in E-Commerce: A Study of Amazon’s Fulfillment Center Network,” which was co-authored with Cornell’s Jean-Francois Houde and Penn State’s Peter Newberry. .... "

Sunday, July 15, 2018

Windows Allies itself with Alexa

Alexa for PC invades your notebook, signs a truce with Cortana in DigitalTrends
Alexa on the PC.  Competing with Cortana,  or just connecting to the increasing number of devices in the home and even office that new are working with Alexa?  Will be fun to see how these devices and their environments evolve. ...  In competition or together.  Sharing skills? ... '

Drug Design with Machine Learning

More on drug design and machine learning.  New methods being tested at MIT.

MIT researchers automate drug design with machine learning
Their model can generate molecules that could be used for therapeutics.
Mallory Locklear, @mallorylocklear in Engadget

Developing and improving medications is typically a long and very involved process. Chemists build and tweak molecules, sometimes aiming to create a new treatment for a specific disease or symptom, other times working to improve a drug that already exists. But it takes a lot of time and a lot of expert knowledge, and attempts often end with a drug that doesn't work as hoped. But researchers at MIT are using machine learning to automate this process. "The motivation behind this was to replace the inefficient human modification process of designing molecules with automated iteration and assure the validity of the molecules we generate," Wengong Jin, a PhD student in MIT's Computer Science and Artificial Intelligence Laboratory, said in a statement. ... " ... ' 

Baidu and Intel Team up for Autonomous Driving

More autonomous driving,  Chinese efforts.

Intel and Baidu team up to make autonomous driving even smarter, safer By Stephen Edelstein 

Two of the biggest names in autonomous driving are teaming up. Intel and Baidu announced an agreement that will see Baidu adopt technology and practices from Intel subsidiary Mobileye as part of its own self-driving car programs.

Under the agreement, Baidu will adopt Mobileye’s Responsibility Sensitivity Safety (RSS) model in both its open-source Project Apollo and commercial Apollo Pilot autonomous-car development programs. Baidu will also use Mobileye’s Surround Computer Vision Kit hardware suite as part of an autonomous-driving system that will be marketed in China.

Unveiled in 2017, RSS is designed to imbue self-driving cars with what Intel calls “common sense human-centered concepts of what it means to drive safely,” such as maintaining a safe following distance and understanding that right of way is given, not taken. The goal is to create a standardized definition of safe driving, and a set of protocols that autonomous-driving systems can be measured against to ensure that they are truly safe. Mobileye and Baidu will work together to adapt RSS to Chinese driving styles and road conditions..... " 

Virtualitics Announces Commercial Release

I had reported on this earlier.  We had experimented with immersing decision makers in data. See my tag links below for an image of our experiment. This is finally what looks to be a good commercial implementation of the idea.  Will look to experiment with it.  Now promoting the inclusion of AI/machine learning methods.  And also now desktop interaction?   Probably a good idea before the AR/VR  revolution takes off.  Also includes what they present as 'innovative maps' that work in 3D.

" .. We are excited to officially launch the commercial release of Virtualitics.

As part of this launch we are offering a free trial program. Simply click here and select the "Request a trial" on the form if you are interested in a trial.

While VR will enhance the visualization of high-dimensional data and provide a real-time collaborative place for your team, it is not necessary. Virtualitics is cross-platform and offers all of the sophisticated AI routines and rich 3D visualizations in desktop as well.

Also please see below a link to a video of our CEO Michael presenting Virtualitics at the Fintech Innovation Lab in NY, with an introduction by Blackrock:

Feel feel to contact me directly with any questions.

Paul GearhartHead of Customer Solutions
Virtualitics ... "

Saturday, July 14, 2018

Predicting Drug Side Effects

AI Helps Stanford Computer Scientists Predict Side Effects of Drug Combinations 
Stanford News
By Nathan Collins

Stanford University's Marinka Zitnik, Monica Agrawal, and Jure Leskovec have developed an artificial intelligence system that can predict potential side effects from the use of combinations of drugs. Their Decagon system could potentially inform clinicians' decisions about which drugs to prescribe and help scientists identify better drug combinations for disease treatment. The researchers compiled a massive network defining more than 19,000 human proteins' interactions with each other and with drugs, then designed a deep learning technique to detect patterns in how side effects crop up based on how drugs target different proteins. They designed Decagon to deduce side-effect patterns and anticipate previously unobserved consequences from taking two drugs together. While Decagon currently only considers side effects associated with pairs of drugs, the team wants to broaden their results to include more complex regimens. ... " 

IBM Says Data, Utility Key to Voice

Having been involved in a test of IBM voice systems, this view is interesting.  Its a new channel that many people are using, so marketers must understand it. 

IBM says data and utility are the keys to brand success in voice
Chief digital officer Bob Lord talks about the perceived limitations of AI assistants—and the vast potential.... " 

Friday, July 13, 2018

Skill Development for Prime Day

The latest for Amazon Skills developers, with particular attention paid for how to use skills for up coming Prime day.  Insightful:

How to Prime Your Skill for Prime Day:

Prime Day is one of the biggest shopping days of the year, which means you have a huge opportunity to reach and delight more customers via voice. Follow these best practices to get your Alexa skill ready for prime time. Learn how to design, build, and launch conversational skills that deliver customer value and drive habitual use.  ... " 

And more for developer education and webinars.  Have attended a number of these and they are interesting in both showing where Amazon is, and what they are driving towards regarding the process of dialog to sell.    And putting in an approach for rewarding skill development.

Chemistry Lessons via VR

And also supporting experimentation with a biochemistry simulation environment, say for Pharma research.  Human augmentation.We did some related research, reported on here.

It's Time for a Chemistry Lesson. Put on Your Virtual Reality Goggles.
The New York Times
By Veronique Greenwood

At the University of Bristol in the U.K., researchers have created a virtual reality (VR) environment that allows biochemists studying a molecule to perform simple tasks nearly 10 times faster than on a two-dimensional (2D) screen-based simulation. The new tool allows users to experience the latest information on what scientists know about how molecules move and flex, says University of Bristol's David Glowacki. Anyone with a virtual reality setup can access the new simulation, which runs on Oracle supercomputers. The researchers timed users, both in VR and on computers with a touchscreen or mouse, on three molecule manipulations. For two tasks, users were significantly faster in the VR environment, while the third task took about the same time using either setup, which Glowacki suggests is because the solution was essentially a 2D movement. The findings indicate that VR could allow scientists to learn about molecule movements much more rapidly. In addition, the new tool could allow researchers who are physically separated, such as at pharmaceutical companies or universities, to examine molecules collaboratively and simultaneously, Glowacki says. .... " 

Analytics Data Catalogs, Approaches, not new

Yes, we know this, and just because some call it AI, does not mean we won't have to gather the data consistently and continually to solve real problems. 

Analytics Industrial Revolution- From The Occult to the Ordinary
By  Snehamoy (Sneh) Mukherjee In Linkedin

Senior Director - Delivering Data Science, Big Data, Machine Learning and Analytics projects for Fortune 500 companies

There is a quiet revolution taking place in the Analytics industry that has the potential to completely turn the industry on its head and the way work gets done in this space. Doomsday pundits have already summoned the evil spirit called AI to put an end to the misery of our uneven paychecks, to be replaced by an Universal Basic Income and some of us have reconciled ourselves to that cruel fate. But before the Apocalypse happens, there is another subtle and continuous tectonic movement happening right under our feet, which if gone unnoticed for long, can catch us in a tidal wave of upheaval in the analytics /machine learning industry.

The typical Analytics (often very eruditely rechristened by brilliant marketers and/or the academia as Machine Learning and a lot would break their heads to prove that the two are different) or a Machine Learning project gets delivered in the following atypical manner in most firms: -

·        Data is pulled from one/multiple tables from a database(s) (by someone who may either be from the client side or by the analytics vendor). It may be a onetime data extraction or if data is needed on a periodic basis, an ETL (Extract Transfer Load or in some cases ELT) process is created to do a batch fetching and processing of files (e.g. weekly transaction data from retail stores, monthly/weekly call data in telecom firms, weekly/daily transaction data in banks) .... " 

Machine Learning Data Catalogs

Makes sense, also connecting the data to its actual meaning,  semantic ontologies,  is also good to do in the same place.      It should be a broader aspect of governance.   It is also an important fundamental aspect of interpretability to know where the data is coming from, what its stability and credibility are.

How the Machine Learning Catalogs Stack Up  
Alex Woodie in Datanami

You can’t do anything with data – let alone use it for machine learning – if you don’t know where it is. In the age of big data, this is not a trivial matter. It is also the main driver that’s propelling the rise of machine learning data catalogs, which the analysts at Forrester recently ranked and sorted. Just a word of warning: the name at the top of the list might surprise you.

According to Michelle Goetz’s June 21 Forrester Wave report, the percentage of analytic decision makers managing more than 1 petabyte of data (either structured, semi-structured, or unstructured) has essentially tripled from 2016 to 2017. That rapid growth has exposed all manner of problems in company’s existing data management and analytic endeavors.

Two of the biggest challenges that companies face today, Goetz writes, are gathering and managing data in a governed manner on the one hand, and managing the business processes that surround the data analytics activities on the other.

“For EA [enterprise analytics] professionals, relying on people and manual processes to provision, manage, and govern data simply does not scale,” the Forrester analyst writes. “Enterprises are waking up to this fact and turning to data catalogs to democratize access to data, enable tribal data knowledge to curate information, apply data policies, and activate all data for business value quickly.”  .... " 

Smart Trash Cans Again

We saw a related device and tested it in our lab smart home, it was designed to detect what you threw away, and then reorder that item.  Or it least put it on your shopping list.  It was not reliable.  It depended on accuracy by having items RFID tagged.   Is this now more reliable with item recognition?

Oscar the A.I. trash can sorts your garbage and recyclables

Confused by which items in your trash are recyclable? Oscar the artificial intelligence garbage bin can do the sorting for you — at least that’s the claim by Autonomous, the ergonomic office and gaming furniture company behind Oscar.

If you are intrigued by a smart device that will help sort your garbage, Autonomous plans to a Kickstarter crowdfunding campaign for Oscar starting July 17. ... "

Neurala Building Ways to Deliver Machine Learning

Have not seen or tried this, making training, use and validation of such models is very useful. 

Neurala launches its Brain Builder to speed up neural network data preparation   By Mike Wheatley

Artificial intelligence startup Neurala Inc. wants to help developers create deep learning applications faster by speeding up the process of feeding data to neural networks to train them.

The company is launching a beta of its new Neurala Brain Builder program, which is a software-as-a-service platform that can tag data used in training models more quickly.

Neurala is trying to address the currently rather cumbersome process of training deep neural networks or DNNs. As the company points out in its pitch, today’s AI applications first need to be “trained” using thousands of images or other pieces of data that must first be tagged so that deep learning models can understand their relevance. However, data tagging is a slow process, and its cost is proportional to the time spent tagging, which means it’s also very expensive.

That’s where Neurala’s Brain Builder can help. The company describes it as a “cost-effective, centralized data preparation tool that enables fast and accurate creation of large quantities of data for DNN training.” It comes with AI-powered annotation tools that can dramatically reduce the time it takes to tag data, while also keeping that data private and secure behind firewalls. Once tagged, the data can be downloaded to create DNNs using popular AI frameworks such as Caffe or TensorFlow.  ... "

Thursday, July 12, 2018

Power of Small Independent Teams

We did this in the AI space.   Yes, provided they are not siloed and have connections to the right resources to build and validate their work.   They also have especially good value when connecting to outside the company resources, acting as a translator , and resource pointer.  Empowered is the key word.   More than agile.  That needs work in many companies.

Unleashing the power of small, independent teams
By Oliver Bossert, Alena Kretzberg, and Jürgen Laartz  in McKinsey

 Small, independent teams are the lifeblood of the agile organization. Top executives can unleash them by driving ambition, removing red tape, and helping managers adjust to the new norms.

What does it take to set loose the independent teams that make agile organizations hum? These teams are the organizational units through which agile, project-based work gets done. The typical agile company has several such teams, most composed of a small number of people who have many or all of the skills the team needs to carry out its mission. (Amazon CEO Jeff Bezos contends that a team is too big when it needs more than two pizza pies for lunch.) This multidisciplinary way of composing teams has implications for nearly every business function. Take IT management. Instead of concentrating technology professionals in a central department, agile companies embed software designers and engineers in independent teams, where they can work continually on high-value projects.

While much depends on the actions of the individual team members, senior executives must thoughtfully create the environment in which teams and their managers can thrive. In a nutshell, senior executives must move the company—and themselves—away from outmoded command-and-control behaviors and structures that are ill-suited to today’s rapid digital world. They must redouble efforts to overcome resource inertia and break down silos, because independent teams can’t overcome these bureaucratic challenges on their own. They must direct teams to the best opportunities, arm them with the best people, give them the tools they need to move fast, and oversee their work with a light but consistent touch. These ideas may sound straightforward, but they go overlooked by too many leaders who’ve grown up in more traditional organizations.

This article explores how senior leaders can unleash their companies’ full potential by empowering small teams and supporting their managers, whose roles have been redefined by agile thinking (exhibit). Let’s start with a glimpse of what that looks like in action. .... "

More on Kroger Fashions, Marketplace Experience

Toured and Wandered the Kroger Marketplace experience yesterday.  Impressive, my only argument was against the sheer sizeof it.  Everything seems to be there, but where?  Will their new Edge system include a navigation capability, or have one in their Kroger App?   Last I saw it the App did not include a navigate from here capability.

Kroger shakes up own-brand fashion with one fell swoop
by George Anderson  with additional expert comments, in Retailwire.

Kroger announced on Monday that it has teamed up with Joe Mimran, the designer behind the Club Monaco, Joe Fresh and Pink Tartan labels, to develop a new exclusive apparel brand for the retailer.

The new line, known as Dip, will include men’s, women’s, juniors, kids, and baby collections. The clothing is being introduced this fall at 300 of Kroger’s Marketplace and Fred Meyer supercenter format stores.

“We’ve worked closely with Joe and his team to develop a line of clothing that works for today’s times — easy to buy, easy to wear, and easy to love. Effortless style, every day of the week,” said Robert Clark, Kroger’s senior vice president of merchandising. “Dip will transform our apparel business, further redefining the customer experience through Restock Kroger.” ... '

AI Conversation with Chris Eliasmith

Another good conversation about AI.  I don't always report these but they are all out at the link.

Voices in AI – Episode 58: A Conversation with Chris Eliasmith
By Byron Reese,  Podcast and transcript 

Episode 58 of Voices in AI features host Byron Reese and Chris Eliasmith talking about the brain, the mind, and emergence. Dr. Chris Eliasmith is co-CEO of Applied Brain Research, Inc. and director of the Centre for Theoretical Neuroscience at the University of Waterloo. Professor Eliasmith uses engineering, mathematics and computer modelling to study brain processes that give rise to behaviour. His lab developed the world’s largest functional brain model, Spaun, whose 2.5 million simulated neurons provide insights into the complexities of thought and action. Professor of Philosophy and Engineering, Dr. Eliasmith holds a Canada Research Chair in Theoretical Neuroscience. He has authored or coauthored two books and over 90 publications in philosophy, psychology, neuroscience, computer science, and engineering. In 2015, he won the prestigious NSERC Polayni Award. He has also co-hosted a Discovery channel television show on emerging technologies.

Visit www.VoicesinAI.com to listen to this one-hour podcast or read the full transcript. ... "

Interpretability Testing Examined

Testing machine learning interpretability techniques  In O'Reilly

By Patrick Hall,Navdeep Gill,Lingyao Meng 

The importance of testing your tools, using multiple tools, and seeking consistency across various interpretability techniques.

This post contains excerpts from the report “An Introduction to Machine Learning Interpretability,”    ... Read the full report on O'Reilly's learning platform.

Interpreting machine learning models is a pretty hot topic in data science circles right now. Machine learning models need to be interpretable to enable wider adoption of advanced predictive modeling techniques, to prevent socially discriminatory predictions, to protect against malicious hacking of decisioning systems, and simply because machine learning models affect our work and our lives. Like others in the applied machine learning field, my colleagues and I at H2O.ai have been developing machine learning interpretability software for the past 18 months or so.

We were able to give a summary of applied concerns in the interpretability field in an O’Reilly report earlier this year. What follows here are excerpts of that report, plus some new, bonus material. This post will focus on a few important, but seemingly less often discussed, interpretability issues: the approximate nature of machine learning interpretability techniques, and how to test model explanations. ... "

Wednesday, July 11, 2018

Open Source, Code, Tech and Decentralized Control

Interesting and considerable piece. What is ultimately the meaning of open source?     Also the architecture of data and code and how it is controlled by just a small number of big players.

Open source hasn’t made tech more open  By Nithin Coca, @excinit in Engadget

Democratic ideals have given way to governments and corporate giants.

here are two institutions dominating the top of the tech food chain today. On one side are big tech companies like Google, Facebook, Amazon, Microsoft and Apple, as well as China's big three of Baidu, Alibaba and Tencent. Alongside them are the massively funded, heavily staffed global cyberpowers -- most notably the US, China and Russia -- who are seeking to monitor and control information flows online in the name of national security or political control.

Both are intertwined. Sometimes intimately, as in China, where an Orwellian social credit system is taking shape, and private companies are becoming indistinguishable from the state's apparatus. In the US, tech companies are now the biggest lobbyists and political donors in Washington, while in Russia there is a battle against the message app Telegram. Together, these forces control the vast majority of information that flows online, either through data gathering, surveillance or censorship.

There is an opposition: Small, often bare-budget operations run by hackers, nonprofit activists and volunteers. These open-source, decentralized projects and cooperative alternatives aim to protect user security and provide them greater control of their personal data. Some, like TOR or Signal, aim to encrypt and protect digital communication from the peering eyes of governments and corporations. Others, like Orchid, Dat or Blockchain-based protocols such as Ethereum want to return the web to it's initial, decentralized roots. Whether or not they get more people to adopt their alternatives could determine whether the years-long trend toward greater corporate and governmental control of data will continue. At stake is nothing less than the future of the internet itself. ... " 

AI, Automation and the Future of Work

Succinctly put: what and why should we address these problems with new tech?  We struggled with  the same thing when we started with this in the 90s. We were already using analytical and statistical methods, but if we had to ratchet up the tech, which problems deserved it?  Which were most risky?  We worried less about the labor implications until we proved we could it at all. What resources did we need?  McKinsey as usual states it well.

AI, automation, and the future of work: Ten things to solve for
By James Manyika and Kevin Sneader

As machines increasingly complement human labor in the workplace, we will all need to adjust to reap the benefits.

Automation and artificial intelligence (AI) are transforming businesses and will contribute to economic growth via contributions to productivity. They will also help address “moonshot” societal challenges in areas from health to climate change.

At the same time, these technologies will transform the nature of work and the workplace itself. Machines will be able to carry out more of the tasks done by humans, complement the work that humans do, and even perform some tasks that go beyond what humans can do. As a result, some occupations will decline, others will grow, and many more will change.

While we believe there will be enough work to go around (barring extreme scenarios), society will need to grapple with significant workforce transitions and dislocation. Workers will need to acquire new skills and adapt to the increasingly capable machines alongside them in the workplace. They may have to move from declining occupations to growing and, in some cases, new occupations.

This executive briefing, which draws on the latest research from the McKinsey Global Institute, examines both the promise and the challenge of automation and AI in the workplace and outlines some of the critical issues that policy makers, companies, and individuals will need to solve for.

Accelerating progress in AI and automation is creating opportunities for businesses, the economy, and society

How AI and automation will affect work

Key workforce transitions and challenges

Ten things to solve for  .... "

Augmented Reality in e-Commerce

Its here, useful, especially in smart phone delivered Apps, but I still don't see it as ubiquitous except in some specific contexts.   A survey of examples:

How augmented reality is transforming e-commerce
ugmented and mixed reality are revolutionising industries and allowing brands and businesses to create new immersive experiences for customers

E-commerce is the fastest growing retail market in Europe with sales expected to surpass £215 billion by the end of the year, according to the Centre for Retail Research.

In the UK alone, online sales hit £60 billion at the end of 2016, up 14.9% from the previous year.

As consumers continue to become more comfortable shopping online, retailers and brands are being forced to investigate ever more innovative ways to get their attention.

Luckily, new technologies like mixed, augmented and virtual reality are offering an entirely new shopping experience that is gradually beginning to grab the attention of consumers and brands alike.

Mixed and augmented reality, in particular, are proving effective in allowing consumers to engage with brands and interact with products like never before.

Using the home décor market as an example, machine learning and mixed reality platforms are allowing consumers to virtually redecorate any room in their home before spending any money. ... "

Walmart Testing Cashierless

Our future, an expectation?

Inside Walmart’s journey to cashierless retail
 by Suman Bhattacharyya in DigiDay

Walmart is testing the waters of cashier automation, first by letting customers scan and pay for items within an app, and now, giving in-store reps the ability to help customers pay on mobile devices.

It’s also studying customer comfort levels with automated payments. At this point, it’s not ready to get rid of cashiers — they’ll just be part of a bigger menu of customer checkout choices, according to Walmart rep Ragan Dickens.

Dickens would not comment on whether fully automated stores are in its plans. But for the past two months, it’s been testing an Apple-store type concept called “Check out with Me” at 350 stores. It gives the agent the ability to check out a customer on a mobile device. Once the transaction is completed, a copy of the digital receipt is sent to the customer via email or text. Dickens said early results are encouraging, and the retailer plans to roll out this service to more stores.

Walmart’s experiment with combining the physical and digital retail experience is an example of how a large, legacy retailer is combining the strength of a long-established physical presence combined with tech innovations — a hybrid model that may fit the needs of a broad swath of customers with varying levels of comfort with technology. Its tests to reduce dependence on traditional checkout counters center around two models: scan and go, which requires the customer to download an app or use a store-provided mobile device to self-scan items, and an Apple store-type model which puts the cashier capability in the hands of roving employees, who, with mobile devices in hand, complete checkouts for customers. ... " 

The Logic Of Logistics

I was reminded of this book through a recent modeling effort.  A very complete look at classic supply chain problems. Not a how to, or pointer to codes to use, and technical, but very complete survey of  useful supply chain optimization and architecture issues.   We used to be a part of the industry group that sponsored some of this work.   Check for the latest edition.

The Logic of Logistics: Theory, Algorithms, and Applications for Logistics and Supply Chain Management (Springer Series in Operations Research and Financial Engineering)  Nov 24, 2004
by David Simchi-Levi and Xin Chen  (MIT of and U of Illinois)

Fierce competition in today's global market provides a powerful motivation for developing ever more sophisticated logistics systems. This book, written for the logistics manager and researcher, presents a survey of the modern theory and application of logistics. The goal of the book is to present the state-of-the-art in the science of logistics management. As a result, the authors have written a timely and authoritative survey of this field that many practitioners and researchers will find makes an invaluable companion to their work.  ... "

End to End Supply Chain talk

Upcoming IBM talk on The combination of AI and Blockchain is interesting.   Attending.

" .. In the age of the start-up disruptors, existing businesses are challenged to change their models to disrupt the disruptors. These changes have begun to focus on supply chain and developing a stronger capability to serve more end customers faster, with less resource tied up in working capital. IBM is answering this challenge for our clients by combining blockchain, AI and B2B EDI to improve data quality, make business predictions, reduce the cost of moving from yesterday’s EDI to tomorrow’s blockchain networking.

Title: The end-to-end supply chain challenge: Blockchain + AI + B2B = Disrupting the Disruptors
Date: Thursday, July 19, 2018
Time: 03:00 PM Eastern Daylight Time
Duration: 30 minutes


Nichole Mumford
Director of Marketing and Professional Development
Council of Supply Chain Management Professionals (CSCMP)
Nichole Mumford has over 20 years of experience in project management, marketing and communications, and a Master of Science in Integrated Marketing Communications from the Medill School of Journalism at Northwestern University. Nichole joined the Council of Supply Chain Management Professionals (CSCMP) in January 2017, and is responsible for marketing, communication and business development efforts of the company. She manages and oversees media relations, branding, advertising, marketing, communications, website development, and sponsorships.Those who work with her appreciate her strong focus on creative problem solving and ability to capitalize on business opportunities.

Rob Allan

Program Director, Supply Chain Insights Offering Management
IBM Watson Customer Engagement
Rob Allan is the Program Director for the Watson Supply Chain Insights offering with IBM. He has been with IBM for over 25 years in a wide variety of supply chain transformation assignments including SAP development, i2 implementation, solution architecture consulting, and big data and analytics projects. He is currently leading a team developing a SaaS platform that leverages Watson machine learning to optimize the way clients are alerted and respond to supply chain disruptions. .... "

Generating a voice Driven Westworld Maze Skill for Alexa

Was wondering what could be done beyond the very simplistic interactions in most Alexa skills.  This Adage article show how some more creative things can be done.   With a few details of the construction, but not enough to do something similar.  Still a creative possibility.  Perhaps as a model for constructing conversational interaction in assistants.  Check out the full article.

Westworld: The Maze: An all-Audio feat of conversational AI
By I-Hsien Sherwood.  in AdAge

Working on a condensed, 15-week timeline, 360i and HBO created "Westworld: The Maze," an Amazon Alexa voice skill that lets players control the actions of a "host," an artificial humanoid in the show's Western theme park, on a quest for personhood. Here's how they built an all-audio feat of conversational AI.

Phase 1: Discovery (3 weeks)
In March, the team decided the Alexa project would be an interactive, all-audio game. They nailed down the creative concept, a series of challenges based on the pyramid of consciousness espoused by Anthony Hopkins' "Westworld" character, Robert Ford: memory, improvisation and self-interest.

Phase 2: Game design and writing (4 weeks)
"You have to start with the user experience: How does the game work? Where are all the places I can go? How do I win? How do I die?" says Andrew Hunter, creative director at 360i. There are more than 60 paths through the game and 32 ways to die, including getting shot, rebooted, eaten by cannibals or trampled in a stampede.

Players speak commands in response to prompts from characters in the story, who might offer to pour them a drink or ask where they would like to go. The writers had to account for anything a player might say, whether right, wrong or nonsensical. ... " 

Tuesday, July 10, 2018

Connecting AI to Mind Mapping

Plan to attend.  Some of the description below is what we tried to use as a method of building a knowledge map that described interaction and process.

How AI Will Transform Mind Mapping?

Webinar:  Date: 17 July 2018 09:30-10:30 (US Central Time) Duration: 60 mins.  Requires Registration 

Artificial Intelligence is poised to have a major impact on many areas of business. Chuck Frey from the Mind Mapping Software Blog believes it will radically alter the ways we interact with mind mapping software. For his new report, Chuck interviewed the software developers in this niche as well as a South American firm that is already using AI to semantically analyze large amounts of text and parse it into mind maps. In this presentation, Chuck shares his findings and provides a tantalizing view of what’s next. .... "

GDPR and Better Marketing

Intriguing point,  if we fully understood the GDPR fully in its application as yet. 

How GDPR Can Lead to Better Personalized Marketing

By Michelle Chiantera in the Cisco Blog

As marketers, we’ve come a long way in becoming data-driven experts when it comes to creating more engaging personalized marketing experiences. The GDPR is making us think more critically about how we leverage that data to be in compliance, while still being able to effectively market and reach customers in relevant ways.

According to Forbes Insights, when it comes to GDPR compliance, 60% of organizations indicated they are challenged with shifting marketing and sales tactics in accordance with GDPR guidelines.

This shows that the GDPR’s impact has been a wake-up call for companies across the board and how they leverage data, but I feel it’s a refreshing and much needed one. Data has been foundational to everything we do for the past few years, but the GDPR is a good reminder that customers are more than targets and data sets, they are people who value trust and partnership.

So how can GDPR make us better at personalized marketing?  ... " 

Kroger Goes Fashion

About to get a tour of a local Kroger Marketplace.  Impressive new additions, including in-store restaurants.   The fashion angle was new to me.

Kroger set to roll out Dip fashion brand
Exclusive apparel label to debut in Fred Meyer, Marketplace this fall   By Russell Redman  

The Kroger Co. has unveiled the apparel private brand it plans to launch in the fall.
Called Dip, the exclusive clothing line was developed by fashion designer Joe Mimran, left, and is slated to roll out to more than 300 Fred Meyer and Kroger Marketplace large-format stores, Kroger said Monday.

Dip will include apparel for women, men, juniors, children and babies. Kroger said the apparel brand is intended to “help busy, on-the-go people live with style and get the most out of their fashion dollar” by offering flexible collections that make it easy for customers to assemble outfits or clothe an entire family. Plans call for the Dip brand to be presented in its own in-store department with the banner prominently featured, according to a conceptual drawing from Kroger. ... " 

Survey: Retailers Missing the Mark

Results of a survey:

How Retailers are Missing the Mark with Shoppers    by Sara Spivey in ChiefMarketer

Consumers are continuously changing the way they shop. They have high expectations for speed and convenience, they consult social media for ideas and inspiration and they’ve become very comfortable making transactions on mobile.

These dynamic shifts in consumer behavior are putting increasing pressure on brands and retailers to create the most innovative shopping experiences to stay competitive, attract new shoppers and foster loyalty among their existing customers. It’s already difficult to identify the shopping experiences that boost optimal sales and revenue, but it’s even trickier to understand whether consumers actually like or want these experiences in the first place.

Sixty-five percent of shoppers don’t consider voice assistants important to the customer experience.
To better understand how retailers are meeting customer expectations and more importantly, the areas where they are missing the mark, Bazaarvoice surveyed more than 400 brands and retailers and 2,000 consumers across the U.S., UK, France and Germany to identify the gaps between what shoppers want and what the industry is delivering. Here’s what we found:

Incorporating the right digital in-store experiences

Innovation and developments in virtual reality and augmented reality continue to make headlines and generate buzz, but blending digital experiences with physical retail is complicated. Though new technologies are flashy and impressive, it begs the question whether they enhance the customer experience or detract from it. Some retailers have introduced creative virtual reality features into their shopping experiences, like the ability to virtually try on outfits or visualize furniture in the home, but are consumers adopting and enjoying these new functionalities? .... " 

Wharton Customer Analytics Accelerator Challenge

FYI much more detail at the link.


Data problems? We can help.

The Wharton Customer Analytics Initiative is looking to provide implementable solutions to companies who are ready to bring their most pressing marketing and analytics problems, and a committed point person, “to the table.”  WCAI wants to help push your company to the next level of critical decision-making, informed by your customer data and the Initiative’s deep knowledge of best practices in customer analytics methods.  Whether you have data to analyze or are in need of a road map to begin, WCAI wants to hear from you.

The 2018 Analytics Accelerator Challenge company application will open on Wednesday, August 15th.  To learn about last year’s Challenge and Summit click here.    ... "

Digital Persons Emerge Again

Recall our long look at using digital personas to represent brand equities, including integrating a chatbot to represent useful information to the consumer.   Our consumers reacted well to this, but the interest faded.  That  also include a personality and image that places her in the 'Uncanny Valley'. 

ANZ has birthed its first "digital person". Jamie, an AI invention styled as a 25-year-old New Zealand woman, started work on Tuesday morning.

Jamie's first job is to chat with customers on the 30 questions the bank gets asked most often by customers.  Though "she" is capable of learning, it's "moderated" learning, so customers trying to teach Jamie swear words, or bad habits, will fail.

But chatting with Jamie for a few minutes results in her dropping hints that there's something more to her than a series of rote answers.  Ask her about her weekend, and she may tell you she enjoys ice dance. ... " 

Monday, July 09, 2018

New Parallel Sampling Algorithm for Speed

Nice to see all of the algorithm research does not have to be AI based.    Note the adaptive sampling approach.  A sort of computationally-crowd-sourced idea?   Applications?   Authors: Singer and Balkanski

Breakthrough' Algorithm Exponentially Faster Than Any Previous One 
John A. Paulson School of Engineering and Applied Sciences
By Leah Burrows

Researchers at Harvard University's John A. Paulson School of Engineering and Applied Sciences (SEAS) have developed a new algorithm that exponentially speeds computation by dramatically reducing the number of parallel steps required to reach a solution. The new approach allows the researchers to speed computation for an enormously large class of problems across many different fields, including computer vision, information retrieval, network analysis, computational biology, and auction design, among others. "We can now perform computations in just a few seconds that would have previously taken weeks or months," says SEAS researcher Yaron Singer. The new algorithm samples a variety of directions in parallel; based on that sample, the program discards low-value directions and chooses the most valuable ones to progress towards a solution. The researcher was presented at the ACM Symposium on Theory of Computing, June 25-29 in Los Angeles, CA.  ... " 

Also mentioned in IEEE Spectrum with examples:

New Optimization Algorithm Exponentially Speeds Computation
Finding the optimal solutions to complex problems can be dramatically faster  .. By Charles Q. Choi ... " 

Potential AI Disruption in Industries

A good set of examples in industries.  Not very much detail, in fact speculation as to direct use in many cases.  Still useful to think about their implications.  Even if AI, in the sense of Machine Learning,  cannot address these cases, it shows a vulnerability that is still there for other technology.    Remember that digitization and analytics by itself can be disruptive.

A snapshot of AI disruption in 27 industries by Entefy

A report from McKinsey estimated that as much as $12 billion was invested globally in artificial intelligence technologies during 2016, including projects focused on machine learning, natural language processing, computer vision, and autonomous vehicles. That figure cuts across industries, and in real-world terms represents thousands of individual R&D projects. 

Unless you’re closely monitoring developments in artificial intelligence, you probably learn about new AI technologies one headline at a time. “New AI system beats champion Go player” or “Advanced AI improves doctors’ diagnoses.” That sort of thing. 

But you can get a much better sense of the scope and diversity of newly emerging AI technologies by learning about a lot of them at once. In this case, 27 different projects transforming 27 different industries. Which is what our research has produced, as you’ll see below.

News of a single AI technology can be pretty exciting, even inspiring. Understanding the diversity and vision of dozens of them underscores just how powerfully transformative this current generation of AI technologies already is, or will soon be.

Here’s our roundup of disruptive AI technologies being designed, built, or deployed in 27 different industries: .... "   (List of 27 industries follows) 

Amazon Moving on Pharma

The next step, the next category, in this case a very big one.  With expert comments.

Amazon lowballs CVS and Walgreens on OTC med prices    by George Anderson in Retailwire

A comparison of private label over-the-counter (OTC) remedies sold by CVS and Walgreens with those from Amazon.com finds that the brick and mortar pharmacy giants are charging substantially more than the e-tailer on like items.

A new report by Jefferies Group found that CVS’s prices were 20 percent higher than Amazon’s and that Walgreens sold its private labels at a 22 percent premium to the e-tail giant, which recently announced that it was acquiring the online pharmacy PillPack.  The acquisition promises to put Amazon in direct competition with CVS and Walgreens for share of the prescription medicine market. PillPack, which delivers pre-sorted doses of prescribed medicines in envelopes, is licensed to fill prescriptions in all 50 states. ... " 

Another look at the IFTTT Platform

Remain intrigued by the broader potential for using IFTTT as a broader intelligence platform.  Anyone looked at that?  An editable rule base?  Driven learning from IFTTT data?  What would it take to test?    Have used IFTTT since its inception.  Here is a new overview.  The type and breadth of clients is interesting.    Easy to set up a test of the idea with a broad view of data. Here is their latest overview:

IFTTT Platform
One connection, 
countless possibilities ... 

Try now
There’s no need to build 3rd party integrations in-house when you can integrate with IFTTT.
Give your users immediate access to over 600 apps, devices, and brands  ... " 

AI Governance Strategy

Not much detail here, but governance in this space is important,  regarding the data and results from the analyses.    And how it relates to ongoing maintenance of solutions.  In particular we found maintenance one of the biggest problems in delivering AI style approaches.  Further the inherent risk involved for any solution has to be modeled and tracked.  Start with an understood process model and build from there.

Opinion: How organizations can develop an AI governance strategy   By  Vikram Mahidhar

Today, many companies are entrusting their top business-critical operations and decisions to artificial intelligence. Rather than traditional, rule-based programming, users now have the ability to provide machine data, define outcomes, and let it create its own algorithms and provide recommendations to the business. .... " 

Buying By Voice Hard Sell for Consumers?

Thoughtful piece.   Yes, it does create a nervousness about making a mistake.  But if enough capabilities are included to attest to the credibility of a request, say with a pin number, and clear repetition of the proposed purchase, and a easy way to revoke the purchase are included in the interaction, people do get used to it.    Voice use in business is also increasing, making people yet more comfortable with the idea.

Shopping by Voice a Hard Sell for Consumers  in Strategy-Business

By Denise Dahlhoff, research director of the Baker Retailing Center at the Wharton School of the University of Pennsylvania. ... 

AI-based voice communication has finally caught on, seven years after Apple’s Siri launch marked the first major step toward making it part of our daily lives. The recent introduction of Amazon’s Echo Dot Kids Edition was the signal for me that voice technology is here to stay. Children are growing up talking to humanized devices as if it’s the most normal thing. My nieces matter-of-factly interact with Amazon’s Alexa all the time, getting “her” to play songs for their kitchen dance shows, checking math homework solutions, and asking Alexa fun questions. (“Are you a girl? Are you bold?”) When the 7-year-old learned that Alexa can shop, she jokingly said, “Alexa, buy me a TV.”

The idea of actually using Alexa to buy something is an unusual concept to her, which seems to be the case with much of the adult population, too. As of now, voice assistants are mainly used to access news, weather updates, and other information, and to play music and podcasts. But voice shopping hasn’t taken off. It is like a magic carpet with the potential to transport us instantly and conveniently to any shopping destination we can imagine, but part of the carpet is still stuck to the ground. The infrastructure of devices that support voice-assisted shopping has greatly expanded, but software components are underdeveloped. .... " 

Sunday, July 08, 2018

Duplex AI and the Call Center

Google is for now denying that Duplex will take over the enterprise call center.  And you should ask if just sounding and acting like a human is enough to make such a chatbot more valuable.   As I have often said, having a memory, an architecture for solving problems in context,  a means to adapt to the interaction and continuous updating from history,  are probably most useful.  And of course, also handing off a problem to any agent, human or otherwise, who is more likely to solve the problem.    Fooling the caller into thinking the chat agent is human is not by itself very useful.  Google, please do keep working on it, and place it on your assistants to test.

Google’s Duplex AI could kill the call center in Quartz by Dave Gershgorn

The robots on the other side of the customer support line could soon start to sound a lot more human.

Google is reportedly shopping its Duplex AI system around as a tool for call centers, according to The Information, including a large insurance company.

Duplex would handle simple calls for the insurance company, and if the customer started asking complex questions the bot can’t handle a human would step in, according to the report. However, it’s unlikely that AI research will cease after mastering simple conversations, meaning call centers could one day be largely automated using this technology. ... " 

Talk: System for Generating New Service and Business Models

Please join us for the next ISSIP Digital Transformation Speaker Series (see details below, or click here).

via Sorin Ciornei, Series Chair, Cisco Systems
Edu4Inno: A service system for generating new services and corresponding business models

Dr. Christoph Peters, Project Manager, Postdoctoral Researcher, University of St.Gallen, University of Kassel

When:  Wednesday, July 11, 4:30pm - 5:15pm (UTC+01:00) Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna   (10:30 AM ET)

Dr. Christoph Peters is a project manager and postdoctoral researcher at the Institute of Information Management (IWI-HSG) at the University of St.Gallen in Switzerland and the Interdisciplinary Research Center for Information System Design (ITeG) at the University of Kassel in Germany. Christoph studied Business Informatics at the University of Mannheim, Germany (M.Sc.-equivalent) and at the Queensland University of Technology in Brisbane, Australia. Christoph holds a doctoral degree from the University of Kassel, Germany (Dr. rer. pol.). He coordinates several research projects and heads a research group of eight doctoral candidates.
His research focuses on service innovation, the systematic design and management of services and work systems, their digitization and respective business models.

Task Description:
Edu4Inno: An service system for generating new services and corresponding business models” presents an educational service system that enables students to learn about service innovations and service systems from a theoretical background on the one hand and on the other hand, to systematically design new services and corresponding business models. Therefore, partners from practice present their real-world challenge at the beginning of a semester and commit themselves for some interview and interaction time slots. The students work in teams and with the partner while trying to master this challenge by systematically designing new and innovative services as well as corresponding business models. Thereby, each team works out predefined artefacts such as a service model, a functioning prototype, a business model, etc. The elaborated service innovations are pitched in a startup environment in front of the practice partners at the end of the semester. We consider our submission and the overall “Edu4Inno” concept as a highly impactful service innovation that continuously “produces” service innovation competences (for students) and service innovations (for practice). The “Edu4Inno” concept has been implemented several times in the last 3 years at two universities leading to an enthusiastic and innovative atmosphere among students as well as to award-winning service innovations.

During this talk Dr. Christoph will present Edu4Inno and key results from its last years' implementations.

Date and Time : July 11 2018 - 4:30pm (UTC+01:00) Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna     (10:30 AM ET)

Zoom meeting Link:   https://zoom.us/j/984551771

Zoom Callin: (415) 762-9988 or (646) 568-7788 Meeting id 984551771
Zoom International Numbers: https://zoom.us/zoomconference

Thinking about Machine Learning

Good thoughts, always read this, worth subscribing to at the link:

Ways to think about machine learning  By Benedict Evans

 I work at Andreessen Horowitz ('a16z'), a venture capital firm in Silicon Valley that invests in technology companies. I try to work out what's going on and what will happen next. 
I write a blog here, do an annual presentation, send out a popular weekly newsletter, talk too fast on podcasts and think aloud on Twitter. 

We're now four or five years into the current explosion of machine learning, and pretty much everyone has heard of it. It's not just that startups are forming every day or that the big tech platform companies are rebuilding themselves around it - everyone outside tech has read the Economist or BusinessWeek cover story, and many big companies have some projects underway. We know this is a Next Big Thing.

Going a step further, we mostly understand what neural networks might be, in theory, and we get that this might be about patterns and data. Machine learning lets us find patterns or structures in data that are implicit and probabilistic (hence ‘inferred’) rather than explicit, that previously only people and not computers could find. They address a class of questions that were previously ‘hard for computers and easy for people’, or, perhaps more usefully, ‘hard for people to describe to computers’. And we’ve seen some cool (or worrying, depending on your perspective) speech and vision demos. 

I don't think, though, that we yet have a settled sense of quite what machine learning means - what it will mean for tech companies or for companies in the broader economy, how to think structurally about what new things it could enable, or what machine learning means for all the rest of us, and what important problems it might actually be able to solve. 

This isn't helped by the term 'artificial intelligence', which tends to end any conversation as soon as it's begun. As soon as we say 'AI', it's as though the black monolith from the beginning of 2001 has appeared, and we all become apes screaming at it and shaking our fists. You can’t analyze ‘AI’.  .... " 

Continued Conversations in Google Assistant

Gets the assistant closer to a true 'conversation'.  Just enabled it and tested, works well.   Still does not have memory from previous conversation, or setting up an ongoing context, like a human would, that's what I want to see, but the direction is good.   Ultimately an intelligent conversation will strongly enable its use in business.  Specifics of continued conversation and examples here.

" ... Now you can ask follow-up questions without
saying "Hey Google" again. Just enable
Continued Conversation in the Google Home app.

Hey Google, turn on the kitchen lights*

How many grams are in a cup of flour?

Set a 12-minute timer ... " 

Legal Sifter

Brought to my attention:  LegalSifter

LegalSifter Essentials

" ... Out-of-the-box contract review, curated by our global network of contract experts
Starting at $25 to $150** USD/user/month
Choose Help Text written by our standard Ken Adams and David Tollen or one of our Combined Intelligence Partners. Get up and running in a few hours. ... "

Smart Contracts that Learn

A very useful next step, examining,

Smart Contracts that Learn   by Michael Slinn in InfoQ
Micronautics Research Corporation

View Presentation, talk

Michael Slinn discusses Smart Contracts, what they are, various implementations, how they can learn, and use cases.  .... "

Saturday, July 07, 2018

Saturday Data Science Reading from DSC

Good Saturday Analytics and Data Science Reading from DSC

Always interesting, at various levels of complexity, join the DSC

Posted by Vincent Granville 
Monday newsletter published by Data Science Central. Previous editions can be found here. The contribution flagged with a + is our selection for the picture of the week. ... "

Times Series with RNN Neural Nets

Been re-examining neural networks for time series models and forecasting.  In the long ago work of modeling with neural nets we had determined it was not useful, but new architectures of recurrent Neural nets RNN make it worth another look.  Here is an examination with Tensorflow

Building Recurrent Neural Networks in Tensorflow

Posted by Ahmet Taspinar  in DSC

Recurrent Neural Nets (RNN) detect features in sequential data (e.g. time-series data). Examples of applications which can be made using RNN’s are anomaly detection in time-series data, classification of ECG and EEG data, stock market prediction, speech recogniton, sentiment analysis, etc.

This is done by unrolling the data into N different copies of itself (if the data consists of N time-steps) .
In this way, the input data at the previous time steps t_n - 1, t_n - 2, t_n - 3, ... , t_0 can be used when the data at timestep t_n is evaluated. If the data at the previous time steps is somehow correlated to the data at the current time step, these correlations are remembered and otherwise they are forgotten.

By unrolling the data, the weights of the Neural Network are shared across all of the time steps, and the RNN can generalize beyond the example seen at the current timestep, and beyond sequences seen in the training set. .... "

Talking about and Measuring Emotions

Way back when we worked with MIT on measuring emotions when people interacted with our products, we discovered the difficulty of measuring this consistently.   Later when we looked at 'neuromarketing' methods, things were not much better.   A number of machine learning methods now claim to make this easy.  Consideration of some derived sentiment based on emotion.  A neuroscientist looks at the issue of naming emotions:

A neuroscientist explains why we need better ways to talk about emotions
We still don’t know what an emotion is
By Angela Chen   @chengela in the Verge  ... " 

See also, Sentiment Analysis.

GoToMeeting Links to Alexa, Voice Transcription

Like to see the closer integration between voice intelligence and meetings.  A very obvious place to improve meeting efficiency and creating links for meeting supporting knowledge.    But all this needs to be very transparent and error free for it to work for business.

GoToMeeting adds AI transcription features, Amazon Alexa integration
LogMeIn-owned videoconferencing tool also gets text chat function. 
By Matthew Finnegan

LogMeIn has updated its GoToMeeting video and audio conference platform with new features that include a text chat function, AI transcription service and integration with Amazon’s Alexa voice assistant.

LogMeIn — which also owns cloud meetings tool join.me — acquired GoTo products from Citrix for $1.8 billion in 2016. This immediately positioned the vendor as one of the largest video and web conferencing players, with millions of customers worldwide. .... "

Friday, July 06, 2018

Autonomous Trucks in Logistics Centers

Makes lots of sense to experiment with this kind of approach where you have the most control.

Autonomous Trucks for Logistics Centers
By Fraunhofer-Gesellschaft 

The Fraunhofer Institute for Transportation and Infrastructure Systems IVI (Fraunhofer IVI) in Germany is collaborating with industry to develop autonomous truck technologies for logistics centers.

The core of the AutoTruck project is the HelyOS "highly efficient online yard operating system," which can be operated via Internet browser. The system displays trucks on a digital map, enabling operators to not only view where individual vehicles are, but also to track them and acquire status information, and transmit missions and work orders to them using the TruckTrix live maneuver planning system.

TruckTrix calculates the trajectory along which the truck is to travel, accounting for vehicle geometry, fixed obstacles, and routes of other driverless vehicles. The calculated paths are sent to the trucks, which are guided by algorithms controlling each vehicle's speed and direction. ... " 

On AI and the Erosion of Reality

This gets back to  the whole question of transparency.  Yes humans have long been manipulating this.   The drawings on the cave walls to start with.   I remember my earliest days in marketing asking the question, what is the truth once you have spent millions in marketing a product?  The question will always be there.

The Erosion of Reality  in SciAM
The most immediate AI threat may be the distortion of truth; something we, and other species, have been doing for a long time.     By Caleb A. Scharf

Microsoft and Marks & Spenser

Not many details, but another example of retail likely using advanced machine learning pattern recognition approaches.  Particular mention of optimizing operations, which would imply process, but not sure whose words these are.

Microsoft, British retail giant in artificial intelligence deal  By Marianne Wilson in ChainstoreAge

Marks & Spencer Group is exploring artificial intelligence through a strategic partnership with Microsoft Corp.

The collaboration is focused on testing the integration of Microsoft AI technologies into M&S’ customer experience, stores and wider operations. The two companies will work together to explore how artificial intelligence can be used the retail environment to improve customer experience and optimize operations.  ... "

Penn on Blockchains

A long 12 page article in my alumni Mag on Blockchains.  With considerable emphasis on 'Smart Contracts'.   Uses in Medicine and Law and elsewhere.  Quite non technical.  Could be good for executive browsing on the topic, with some useful hype cautions.

Blockchain Fever in the Pennsylvania Gazette  by Trey Popp
Cryptographic sorcery, entrepreneurial zeal, and utopian dreams have gripped a striking number of Penn students and alumni this year. Why are people so excited?  ... "

A Guide to Ensemble Learning, with my Cautions

Nicely done.  Ensemble methodology means using a number of different solution methods to solve a problem, and finding the best individual or combination of methods.  Caution must be taken to ensure that the assumptions for each method is reasonable, that the right metadata is available to apply the solution, and the solution is stable under its assumptions. 

Often this means that the model needs to be simplified to some degree.  Make sure the simplification is reasonable for purpose and that owners of the data and process agree.   Have seen reasonability slip away under pressure for a best solution.

Guide to Ensemble Learning with Python code, considerable depth, via O'Reilly

A Comprehensive Guide to Ensemble Learning (with Python codes)
By  Aishwarya Singh in Vidhya Learning .... 

Thursday, July 05, 2018

Was there a GE Model, and What Killed it?

I was impressed by what I saw at GE in a number of interactions, they have been mentioned in this blog from a tech perspective a number of times.  Is it just because the tech has not yet clicked?  This piece was enlightening to me.  Sharing capabilities in tech requires setting good standards.  Saw this, but not universally.   In the newer still emergent tech they are doing very well to provide goals and paths to achieve them.

Who Killed the GE Model?
By Benjamin Gomes-Casseres in the HBR

Who killed GE?

Of course, GE is not dead, and it may well revive and flourish as a company. After all, IBM came back from the dead in the 1990s. But the GE model is dead — and there’s a long list of possible suspects.

The GE conglomerate combined a wide range of industrial businesses under one roof. Unlike a pure holding company or a modern hedge fund, the GE model intended to create value by actively sharing capabilities among its disparate businesses, which, with one important exception, were all rooted in manufacturing. .... " 

Uncovering the Details

Good, detailed piece in the Tableau blog.   Have used Tableau in two large enterprises, and its a great tool.  But I further like the broader implication of the title.  Analytics of any level of sophistication and 'intelligence' can uncover details.  What we sometimes disparagingly called 'descriptive' analytics.   Working with your data closely, with people that know the data and where it comes from and is used, can lead to valuable insight.   Even before you get predictive or prescriptive.  Keeping it simple can work.

How analytics helps cash flow management by uncovering key details
By Ed Barrie Senior Director, Treasury at Tableau ....

More Smaller, Multi Environment Robotics

Out of Harvard.   Some of our earliest looks at robotics were where they might be used in varying environments.  So this is always interesting.  Expect them to continue to shrink.  Includes video of this weird looking device.

Harvard's tiny robot can swim and walk underwater
The Ambulatory Microrobot is at home on land or a (very small) sea.

By Jon Fingas, @jonfingas in Engadget ....

Demystifying Data Science Free online Conference.

Looks to be of interest with good industry and technical speakers.

A FREE Live Online Conference for Aspiring Data Scientists & Data-Curious Business Leaders
28 Speakers • 2 Days • FREE
July 24 - July 25, 2018     10am - 5pm ET

The #DemystifyDS Experience
Demystifying Data Science is designed to be equal parts informative and interactive. All registrants will have access to the presentation recordings after the conference - but you have to attend live for the full experience!

More information and registration.

Neural Compute Stick in USB Form

Considering how this would be useful for training, delivery on some devices.  For taking an update to a IOT device.

Intel NCSM2450.DK1 Movidius Neural Compute Stick
4.8 out of 5 stars    12 customer reviews  | 7 answered questions   
Price: $76.05 
Neural Network Accelerator in USB Stick Form Factor
Real-time on-device inference; no cloud connectivity required
No additional heat-sink, no fan, no cables, no additional power supply
Prototype, tune, validate and deploy deep neural networks at the edge  ... " 

Tuesday, July 03, 2018

Factor Analysis and Bayesian Networks

Factor analysis was a favorite statistical method we used to analyze complex influences.  Here is a link to a Bayesian approach. 

Factor Analysis Reinvented—Probabilistic Latent Factor Induction with Bayesian Networks and BayesiaLab

Bayesian networks have been gaining prominence among scientists over the last decade, and insights generated with this new paradigm can now be found in books and papers that circulate well beyond the academic community. Practitioners and managerial decision-makers see references to Bayesian networks in studies ranging from biostatistics to marketing analytics. Therefore, it is not surprising that the relatively new Bayesian network framework prompts comparisons with more conventional methods, such as Factor Analysis, which remains widely used in many fields of study.

The goal of this webinar is to compare a traditional statistical factor analysis with BayesiaLab's new workflow for Probabilistic Latent Factor Induction using a psychometric example.

More information, slides, presentation.

What to do with Python

Useful starter examples.

What to do with Python 
By YK Sugi   Founder at CS Dojo, formerly @ Google

What exactly can you do with Python? Here are Python’s 3 main applications.
If you’re thinking of learning Python — or if you recently started learning it — you may be asking yourself:

“What exactly can I use Python for?”
Well that’s a tricky question to answer, because there are so many applications for Python.

But over time, I have observed that there are 3 main popular applications for Python:

Web Development

Data Science — including machine learning, data analysis, and data visualization


Let’s talk about each of them in turn.  .... 

Wal-Mart Introduces 3D Online Shopping

Walmart kicks off back-to-school season with 3D online shopping
By Deena M. Amato-McCoy in ChainstoreAge

A discount giant is making it easier for co-eds to decorate their dorm rooms.

Walmart is enhancing its new website with two new services that will streamline how customers browse and make purchases across its home furnishings offering. On Thursday, the discount giant will begin testing its 3D virtual shopping tour, a service that enables customers to virtually browse a curated apartment. The service features nearly 70 items from both national brands and Walmart’s private label offerings. ... " 

Polinode Adds New Influencer and Collaboration Features

Polinode has been frequently mentioned here, noe new features of interest.   You can easily test it.

Introducing an Integrated Tool to Identify Influencers and Dynamic Collaboration Matrices  By Andrew Pitts

If you haven’t used Polinode before - it’s a tool for mapping, visualizing and analysing networks in the browser. With Polinode you can either create Networks, Surveys or both. Polinode Networks allow you to upload any type of connected data to the cloud and then visualize and analyze this data directly from your browser. The source of the data can literally be anything including emails, 360 degree performance reviews, enterprise social networks, social media, etc. 

Polinode Surveys allow you to collect network data using our built-in relationship-based survey tool. For example, you can ask questions such as “Who do you work with often?” or “Who do you go to for advice?”. You can then visualize and analyze this network data in one integrated and highly flexible solution. Applications include change management, identifying emerging talent, M&A integration and improving workplace layouts. This is what we call the ‘Survey’ functionality and in summary it is most often used for organizational network analysis (or ONA for short).

We are excited to announce that two new powerful features have just gone live! A handful of users have been testing these features over the last few weeks and the feedback has been very positive so we wanted to share them with the broader community as soon as possible. In summary, these new features are:

Identify Influencers: We have added the ability to identify those individuals or nodes in the network that together maximise the coverage over the entire network. 

Collaboration Matrices: We have added a new type of report called a Collaboration Matrix that summarises the interaction between different groups in a network. .... " (Much more in this post) 

Linking Robotic Process Automation with AI

Have been spending some time looking at how RPA can be derived from business process models.  Its natural to think how AI/machine-Learning could also drive the associated processes from patterns found in existing or proposed process.   Starting with simple screen scraping, and beyond to work flow.  Here is a starting point:

AI is becoming the nucleus of intelligent apps for robotic process automation by James Kobielus in SiliconAngle

Robotic process automation, software that emulates how people carry out tasks in a process, is becoming one of the principal enterprise use cases for artificial intelligence.

Established RPA solution providers are becoming prominent players in the cloud application development arena. Chief companies in this segment — which include Automation Anywhere Inc., BlackLine Inc., Blue Prism Group, Kofax Inc., Pegasystems Inc. and UiPath — are blurring the already fuzzy lines among RPA, business process orchestration, Web content management, edge computing and application development. And most of them now emphasize the depth of their solution portfolios’ AI capabilities.

When we speak of AI in an RPA context, the discussion has traditionally been focused on the technology’s use in inferring an application’s underlying logic from artifacts that are externally accessible at the client level. In this regard, machine learning and other AI tools typically drive the screen scraping of user interface presentation elements, optical character recognition of onscreen text, autosensing of browser-level control and domain object models, capture of human-user keystrokes and clicks, understanding of natural-language text and parsing of document object models  .... " 

Monday, July 02, 2018

(Update) Cognitive Assisted Interactive Labeling

Via: Karolyn Schalk,  Manager
Executive and Technical Expertise, Cloud, AI and IT Operations

CSIG Talk, July 5 10:30 AM ET: Tensorboard Speaker: Francois Luus, IBM

Speaker:   Francois Luus
Title:  “Cognitive-assisted Interactive Labeling & Software 2.0 “

Future cognitive systems will largely make use of self-optimizing cognitive models, called Software 2.0, where the primary remaining task will be to provide labels/supervision. Software 2.0 will provide fully learned programs even for complex objectives and program functions, just requiring the human to specify the desired behavior. Human involvement in machine learning is thus evolving from engineering features, to hyperparameter optimization, and finally to providing supervision. In this talk we propose the use of machine learning itself in the Toolchain for Software 2.0, where the objective is to efficiently obtain and manage supervision. We modify a versatile SGD-based dimensionality reduction algorithm to allow for feature space quality assessment and for direct editing of the feature space itself, with labels as final output for use in image classifier training. We show the improvements in labeling efficiency through cognitive assistance for a variety of benchmark datasets.

Bio:  Francois Luus is a Research Scientist at the Johannesburg lab of IBM Research | Africa where machine learning is advanced and applied in the domains of healthcare, environment and finance. He holds a PhD in Computer Engineering and his research interests include computer vision, deep learning and dimensionality reduction, which led to innovations in Earth Observation during his time at the Council for Scientific and Industrial Research. Previously he conducted information-theoretic coding research at the Sentech Broadband Wireless Multimedia Center aimed at improving the robustness of wireless communications.

Zoom meeting Link: https://zoom.us/j/7371462221
Zoom Callin: (415) 762-9988 or (646) 568-7788 Meeting id 7371462221
Zoom International Numbers: https://zoom.us/zoomconference

(Check the website in case the date or time changes: http://cognitive-science.info/community/weekly-update/    Also access to slides and recording )