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

Wednesday, January 31, 2018

Bain Davos Video: Delivering Value to Customers

Davos 2018: Bain Breakfast  Video and transcript

January 24, 2018 Bain video By Orit Gadiesh, Eric Almquist, Jørgen Vig Knudstorp and Maurice Lévy

Orit Gadiesh, chairman, Bain & Company, introduced three speakers at the 14th annual Bain Breakfast at the World Economic Forum's annual meeting: Eric Almquist, a partner in Bain's Customer Strategy & Marketing practice, Jørgen Vig Knudstorp, executive chairman of Lego, and Maurice Lévy, chairman of the supervisory board at Publicis Groupe. All three speakers covered the topic of delivering value for customers. .... "

Voicera's Eva for Meetings

Just reviewing this capability:  Voicera   Have tried it once, will follow with more.

See also:  https://www.voicera.com/how-it-works/

" ... Make Meetings Actionable.

You focus on the people, we’ll help capture what matters.
Meet Eva, your in-meeting AI assistant.
Eva listens and records the conversation, helps take notes and identify action items to be organized for follow up. Connect what gets said to what gets done.
“It’s only human to forget to take notes. That’s why I invite Eva.”
Raphael Tremblay, Vice President of Sales, TableUp

Automated Note-Taking, Easily Share Your Notes, No Hardware or Software, Just Talk or Tap to Highlight, Eva works for all types of meetings .... " 

Images Floating in the Air

Seems to compete with augmented reality scenarios I have seen.  Do we need a headset?     Advertising and more.  Not a Hologram, which powered the examples of this I have seen.

BYU Study Produces 3D Images That Float in 'Thin Air' 
BYU News (UT)
By Todd Hollingshead

Researchers at Brigham Young University (BYU) say they have developed a method to create a three-dimensional (3D) volumetric projection, or a 3D image that floats in air, that a viewer can walk around and see from every angle. The researchers note they developed a free-space volumetric display platform, based on photophoretic optical trapping, that produces full-color, aerial volumetric images with 10-micron image points by persistence of vision. The new technique uses forces conveyed by a set of near-invisible laser beams to trap a particle of cellulose and heat it evenly. This process enables the researchers to push and pull the cellulose around, while a second set of lasers projects visible light onto the particle, illuminating it as it moves through space. "In simple terms, we're using a laser beam to trap a particle, and then we can steer the laser beam around to move the particle and create the image," says BYU's Erich Nygaard. .... " 

CPG Leader on Voice and e-Commerce

Well known manufacturer looking at the marketing future:

Clorox CMO: Voice and e-commerce will grow faster than marketers expect

In a Q&A with Marketing Dive, Eric Reynolds dishes on why he's bullish on a number of digital technology areas other CPGs have either stepped back from or might be considerably underestimating. .... " 

" ... Clorox is focusing on creating "channel-first" messages that meet consumers where they are, said Chief Marketing Officer Eric Reynolds. "[W]e're getting to a point now where digital is not enough; we have to break it down into its various components," he said. ... " 

No Exceptions

Nicely done piece by William Vorhies of DSC.  Worth always keeping in mind.  He also provides a favorite quote from James Thurber.   I like it.   Good detail at the link:

When Variable Reduction Doesn’t Work
Posted by William Vorhies     

Summary:  Exceptions sometimes make the best rules.  Here’s an example of well accepted variable reduction techniques resulting in an inferior model and a case for dramatically expanding the number of variables we start with.

One of the things that keeps us data scientists on our toes is that the well-established rules-of-thumb don’t always work.  Certainly one of the most well-worn of these rules is the parsimonious model; always seek to create the best model with the fewest variables.  And woe to you who violate this rule.  Your model will over fit, include false random correlations, or at very least will just be judged to be slow and clunky.

Certainly this is a rule I embrace when building models so I was surprised and then delighted to find a well conducted study by Lexis/Nexis that lays out a case where this clearly isn’t true.  ... " 

Amazon Fresh and Fresh Grocery Business

Amazon is dropping major hints it's ready to dominate the fresh grocery business   By Dennis Green in BusinessInsider

The head of Amazon's Prime Now division is now also leading its AmazonFresh grocery business.   Amazon Fresh has scaled back while Prime Now expands to new cities.  Synergies with Whole Foods and Prime Now may prove a winning combination.

Amazon could be about to make some big changes to its grocery business, which currently consists of two services: AmazonFresh and Prime Now.

AmazonFresh is the company's oldest grocery delivery service, a traditional online option with a $15-a-month membership cost and reserved delivery times.

Prime Now, on the other hand, is Amazon's two-hour delivery offering, which comes free with Prime. On top of a more limited selection of food, it also offers a small selection of products like Echo devices and seasonal items.  ... "

Tuesday, January 30, 2018

Crowd Sourcing vs Algorithm

We actively used Mechanical Turk, so the comparison is interesting, but depends much on the details of the set up involved.

Mechanical Turkers may have out-predicted the most popular crime-predicting algorithm   The Verge by Russell Brandom

Our most sophisticated crime-predicting algorithms may not be as good as we thought. A study published today in Science Advances takes a look at the popular COMPAS algorithm — used to assess the likelihood that a given defendant will reoffend — and finds the algorithm is no more accurate than the average person’s guess. If the findings hold, they would be a black eye for sentencing algorithms in general, indicating we may simply not have the tools to accurately predict whether a defendant will commit further crimes.

Developed by Equivant (formerly Northpointe), the COMPAS algorithm examines a defendant’s criminal record alongside a series of other factors to assess how likely they are to be arrested again in the next two years. COMPAS’ risk assessment can then inform a judge’s decisions about bail or even sentencing. If the algorithm is inaccurate, the result could be a longer sentence for an otherwise low-risk defendant, a significant harm for anyone impacted.

Reached by The Verge, Equivant contested the accuracy of the paper in a lengthy statement, calling the work “highly misleading.”  .... "

Dynamic Networks

Not overly technical but thoughtful piece. I like any article that aims to deal with differing contexts over time, rately considered well in analysis.  Even in forecasts it is rare that we treat more than one dependent variable.   Look at it like a network.   This article does not do that, but it surrounds the problem.

Elements of the Theory of Dynamic Networks
By Othon Michail, Paul G. Spirakis 
Communications of the ACM, Vol. 61 No. 2, Page 72

" ... A dynamic network is a network that changes with time. Nature, society, and the modern communications landscape abound with examples. Molecular interactions, chemical reactions, social relationships and interactions in human and animal populations, transportation networks, mobile wireless devices, and robot collectives form only a small subset of the systems whose dynamics can be naturally modeled and analyzed by some sort of dynamic network. Though many of these systems have always existed, it was not until recently the need for a formal treatment that would consider time as an integral part of the network has been identified. Computer science is leading this major shift, mainly driven by the advent of low-cost wireless communication devices and the development of efficient wireless communication protocols. .... " 

Harvard's MilliDelta Robotics

Ultimately the very small may be more important than robots at human scale ...

Harvard's milliDelta Robot Is Tiny and Scary Fast    By Evan Ackerman

Harvard's milliDelta is a millimeter-scale delta robot based on origami-inspired engineering that can reach velocities of 0.45 m/s and accelerations of 215 m/s².

In terms of sheer speed and precision, delta robots are some of the most impressive to watch. They’re also some of the most useful, for the same reasons—you can see them doing pick-and-place tasks in factories of all kinds, far faster than humans can. The delta robots that we’re familiar with are mostly designed as human-replacement devices, but as it turns out, scaling them down makes them even more impressive. In Robert Wood’s Microrobotics Lab at Harvard, researcher Hayley McClintock has designed one of the tiniest delta robots ever. Called milliDelta, it may be small, but it’s one of the fastest moving and most precise robots we’ve ever seen. .... "

(Pics and video at the link)  .... 

On the Risks of Trusting Sensors

Technical piece on the risks beyond the software in the IOT

Inside Risks:  Risks of Trusting the Physics of Sensors
Protecting the Internet of Things with embedded security. 

" .... Conclusion:
Sensors are vulnerable to spoofing by transduction attacks. Cyberphysical systems must cope with analog threats that an adversary could exploit without any special-purpose equipment. Automobiles decide whether to deploy an airbag based on data from accelerometers.   Pacemakers and defibrillators decide whether to emit shocks based on data from cardiac sensors.  It is inevitable and predictable that hackers will try to manipulate sensors to cause havoc.

Autonomous systems making safetycritical decisions should remain safe even when an adversary can exploit physics to influence the output of sensors. The community can reduce these risks by designing sensors to be continuously checkable for security properties and by increasing opportunities for students to master the physics of computer security and principles of embedded security. .... " 

Thinking Machines

Like the broadening of what we sometimes call AI.  I like the idea of 'Cognitive' because it promises less.  Would also like to add other classifiers, say 'Process Intelligence'?

Thinking Machines Going Mainstream 
in SIGNAL Magazine  By George I. Seffers

Experts predict cognitive computing will eventually become normalized as a routine behavioral component in any newer systems. "It will be an expectation of the users that this assistive, interactive, iterative role that it plays within decision making becomes the norm," says Cognitive Computing Consortium co-founder Sue Feldman. She believes more interactive technology will be "able to return answers or graphs or whatever is necessary in an iterative manner with the people who are using it," while also being contextual. Meanwhile, consortium co-founder Hadley Reynolds expects an acceleration in the blurring of boundaries between technological devices and other objects, including apparel and everyday appliances, to the point where they vanish completely. Feldman and Reynolds agree big data will continue to fuel advances in cognitive computing, with a major possibility of a new profession stemming from cognitive computing and artificial intelligence. The consortium co-founders alternately cite ethics and trust as the most pressing remaining challenges for cognitive computing. ... " 

Showrooms replacing Closing Stores

Replacing closing stores with Showroom stores ... ?  Implications for manufacturers?   Ability to do selectively more demonstration of more complex products, or products in context?  Use showrooms to introduce new products?

Showroom Stores
So, you're closing some stores this year. Here’s how to do it right. ....

With further expert opinion in Retailwire:

Monday, January 29, 2018

Algorithmia: Best Practices for Algorithm Design

Have not looked at Algorithmia for a while. It is branded as a marketplace for algorithms. Nice idea, don't know how well it is doing. 

They now have a good blog post on Best Practices for Algorithm Design:

Nicely done, lots of good thoughts on the problems. Includes code snippets.   Should be read.  BUT not a word about any kind of inherent bias in the designs. That is a high level decision, certainly, but probably one of the most important considerations to consider.   The bias may come from the particular context of the data used to derive the algorithm.  So the algorithm also must include all the context,  but I have found that is rarely done.

" ... We host more than 4000 algorithms for over 50k developers. Here is a list of best practices we’ve identified for designing advanced algorithms. We hope this can help you and your team. ... "

What is Algorithmia?

" ... Home
Welcome to the Algorithmia Developer Center. Here you can find guides showing how to use the Algorithmia Clients in the language of your choice, find tutorials for Working with Data on Algorithmia, learn how to Write your own Algorithm or Host your Data Model using frameworks such as Scikit-learn or Tensorflow.   .... " 

Solving NLP Problems

Nicely done, mostly non-technical piece on natural language processing (NLP).  Why it is important and how to plan and execute NLP problems.   Text is our language, still an important channel, so we need to know how to interpret it to get value.  Link to much more ...

How to solve 90% of NLP problems: a step-by-step guide
Using Machine Learning to understand and leverage text.
By Emmanuel AmeisenProgram Director at Insight AI @EmmanuelAmeisen

Text data is everywhere

Whether you are an established company or working to launch a new service, you can always leverage text data to validate, improve, and expand the functionalities of your product. The science of extracting meaning and learning from text data is an active topic of research called Natural Language Processing (NLP).

NLP produces new and exciting results on a daily basis, and is a very large field. However, having worked with hundreds of companies, the Insight team has seen a few key practical applications come up much more frequently than any other:

Identifying different cohorts of users/customers (e.g. predicting churn, lifetime value, product preferences)

Accurately detecting and extracting different categories of feedback (positive and negative reviews/opinions, mentions of particular attributes such as clothing size/fit…)
Classifying text according to intent (e.g. request for basic help, urgent problem)    ... "

Virtualitics for 3D Data Sharing

Took another look at Virtualitics.  The idea is one we experimented with for years and this is a commercial example.  This is a kind of  Virtual Visualization, and particularly attractive if you want
to share visualizations among many people.  Premise is interesting,  easily sharing complex data is good.    Its less clear if the 'immersive' aspects of this capability are very beneficial.   I am on the  newsletter for this effort

Key Features: 

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

Virtualitics Dashboard  :  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) and VR Dashboards
Collaborative and fully customizable shared space where you can analyze data and present and discuss insights. ... "

Humans Gaming the System

Sometimes you need to include outlier training.   But is that an opening for gaming?   Ultimately we will also have to model multiple kinds of human intent as well.

Are AI Learning Scenarios Unpredictable Enough?    By Sam Ransbotham  in MIT Sloan

A fender bender heard around the AI world happened last week in Las Vegas when a self-driving shuttle was involved in a minor collision during its first hour of service. It is ironic that this happened in Vegas, a city based on games. How would you score this match between humans and machines? Is it 1-0 in favor of the humans, a victory for the home team?

Not so fast.

In the aftermath of the “calamity,” sensational headlines played to our default thinking that the machine was to blame. Perhaps we humans reveled a tiny bit in the schadenfreude of seeing our emerging computer overlords beaten so quickly when practice was over and the real game started.

But in this incident, the autonomous electric vehicle was shuttling eight passengers around Las Vegas’ Fremont East entertainment district when a human-operated delivery truck backed into the front bumper of the shuttle. Recognizing the oncoming truck, the shuttle stopped to avoid an accident. The human driving the truck, however, did not stop. We instead need to score this matchup as 0-1 in favor of AI.

Worse, this accident illustrates a crucial challenge in the interplay between AI and humans. Systems are typically configured in contexts without nefarious actors, where players are instead well-intentioned and follow the rules. After all, the first step is to get something working.

How does design for the “well-intentioned” manifest itself here? Consider how the situation unfolded: The shuttle seems to have accurately recognized the situation and predicted an imminent collision. This is a current strength of AI — processing input from many signals quickly to build an accurate short-term estimate of what will happen.

Given the prediction, the next step was more difficult. Should the shuttle have honked? That seems fairly risk-free. Reversed and backed away from the approaching truck? That seems more difficult and riskier than a honk. In this case, the shuttle stopped and did nothing — when in doubt, first do no harm. For imperfect AI, faced with uncertainty, a reasonable default is to stop and do nothing.

But this incident should show businesses that thinking about well-intentioned actors won’t be enough. The first law of robotics doesn’t stop with “a robot may not injure a human being”; it continues with, “or, through inaction, allow a human being to come to harm.”

Now that we know that the shuttle will stop, we have to think nefariously. For example, I live near a busy street, so there is no way that I would step out in front of traffic; there are way too many distracted drivers focused on their mobile devices and not on me. But, now that I know that vehicles will behave like the shuttle, why not step out whenever I feel like crossing the road? I can rely on excellent sensors, prediction, and lightning-fast braking to protect me. Going further, could I create traffic chaos on demand by jumping out unexpectedly? This scenario is not dissimilar to the denial-of-service attacks on computer systems, where attackers can shut down systems and hold them for ransom. .... " 

Sunday, January 28, 2018

Open Source Leadership

The Emperor Has No Clothes: Recasting Leadership in the Open-source Era
Rajeev Peshawaria discusses his new book: Open Source Leadership: Reinventing Management When There's No More Business as Usual.

The digital age has democratized the workplace. Now employees can wield just as much knowledge and voice as their managers. It’s a profound change that is forcing an evolution in leadership. Rajeev Peshawaria, who heads the Iclif Leadership and Governance Centre in Malaysia, explores the idea in his new book, Open Source Leadership: Reinventing Management When There’s No More Business as Usual. He contends that leaders must learn to do things differently if they want their companies to innovate and survive. Peshawaria has held senior leadership positions at American Express, HSBC and Goldman Sachs, and he was chief learning officer at Coca-Cola and Morgan Stanley. He discussed his ideas on the Knowledge@Wharton show, which airs on Wharton Business Radio on SiriusXM channel 111. (Listen to the podcast at the top of this page.)  The following is an edited transcript of the conversation. .... " 

Mycroft: Better Privacy and Open Source Assistant

Wrote about Mycroft back in April 2016,  have been following.  Still says it won't be available until late 2018.   You do have to be careful about funding some of these efforts, there is always a considerable possibility it won't ship ever.  Better privacy and open source are good things, and would like to see an option out there that could be better tailor-able to those concerns.

Mycroft Mark II: The Open Source Answer to Amazon Echo and Google Home That Doesn’t Spy on You   By Abhishek Prakash

Meet Mycroft: An Open Source Virtual Assistant
Thankfully, we have an open source virtual assistant called Mycroft. It doesn’t violate your privacy, it is open and customizable. You can get the source code of Mycroft on GitHub.  .... "

My complete index of assistants.  Continually updated.

Will Amazon Eat Media and Marketing?

Interesting idea posted, where is Amazon going?  A considerable document.  In Digiday: 

" ... Amazon seems to have everything going for it: Endless scale, bottomless coffers and a customer-centric point of view that can provide the kind of data marketers and media have only been dreaming about for the last few years. If Amazon can improve some of its advertising back-end and focus on growing its non-commerce business, 2018 will certainly be the year of Jeff Bezos. ... " 

Novartis as a Data Science Co.

Like to hear how leaders are positioning this.  Is it just about data science as it exists today, or is it more about data science as it supports process?   The latter I would think.

Reimagining Novartis as a 'medicines and data science' company  By Vas Narasimhan

Global Head of Drug Development and Chief Medical Officer at Novartis

Smart Contract Development

Following up on Smart Contracts beyond the Bitcoin.

New Project to Fight Deficit of Smart Contract Developers  By Nick Bakursky in Cointelegraph

The number of job postings on LinkedIn related to Blockchain, cryptocurrency and Bitcoin, increased at least fourfold in 2017, according to data provided by Bloomberg Law. As the source says, the number of such jobs is expected to grow in 2018 as well. Needless to mention, skillful developers are few now and in high demand. A new project titled Fabric Token intends to solve the problem of the lack of Blockchain specialists and current difficulties in decentralized application development by setting up a marketplace and launching the application for smart contract generation. ..."

See FabricToken .... Descriptive white paper of Fabric Token Ecosystem.

See also their blog, good to see they have included BPM expertise in their team.

Saturday, January 27, 2018

Why Digital Strategies Fail

Thoughtful piece on the topic:

Why digital strategies fail
By Jacques Bughin, Tanguy Catlin, Martin Hirt, and Paul Willmott in McKinsey

Most digital strategies don’t reflect how digital is changing economic fundamentals, industry dynamics, or what it means to compete. Companies should watch out for five pitfalls. .... 

We find that a surprisingly large number underestimate the increasing momentum of digitization, the behavioral changes and technology driving it, and, perhaps most of all, the scale of the disruption bearing down on them. Many companies are still locked into strategy-development processes that churn along on annual cycles. Only 8 percent of companies we surveyed recently said their current business model would remain economically viable if their industry keeps digitizing at its current course and speed. ... " 

Human Designed Machine Learning

Good thought on the problem, what are the specifics of human needs, without limiting possible augmentation?  Very good, lengthy piece, mostly about picture taking and related interface.

The UX of AI

Using Google Clips to understand how a human-centered design process elevates artificial intelligence     By Josh Lovejoy in Google Design

As was the case with the mobile revolution, and the web before that, machine learning will cause us to rethink, restructure, and reconsider what’s possible in virtually every experience we build. In the Google UX community, we’ve started an effort called “human-centered machine learning” to help focus and guide that conversation. Using this lens, we look across products to see how machine learning (ML) can stay grounded in human needs while solving for them—in ways that are uniquely possible through ML. Our team at Google works across the company to bring UXers up to speed on core ML concepts, understand how to best integrate ML into the UX utility belt, and ensure we're building ML and AI in inclusive ways. .... " 

Algorithms of Our Future Thinking Machines

More about dealing with uncertainty.    Most of this is today rolled into the statistics of machine learning, but it needs to be included in results that can be easily propagated for reasoning and assisting.   Note especially the need for this in dynamic, contextually rich systems.

The Algorithms of Our Future Thinking Machines 
Uppsala University  By Anneli Bjorkman  

Researchers at Uppsala University (UU) and the KTH Royal Institute of Technology in Sweden are building algorithms for dynamic systems under the NewLEADS project. "Uppsala's focus in this project is to build mathematical models of dynamic systems that can identify and deal with uncertainty," says UU professor Thomas Schon. Among KTH's areas of concentration are designing experiments that extract as much data as possible from a particular system that is to be modeled, notes KTH professor Hakan Hjalmarsson. "I hope...that KTH and UU will be able to build a good theoretical base that is easy to implement in various applications of dynamic systems such as smart climate control of buildings and self-driving vehicles, and that we get to see them put to good use," he says. Schon notes his team's latest area of focus is medical applications, particularly automated diagnosis as a tool for doctors.   .... " 

Wal-Mart to sell Online Books

A more serious move by a retailer into online channels.

Walmart Challenges Amazon with E-Book Move

Walmart plans to sell e-books and audiobooks, its latest effort to encroach on an area of Amazon strength. The nation's largest retailer said Thursday it has struck a deal to use Tokyo-based Rakuten Inc.'s e-reading service called Kobo to sell e-books and audio books. It will also offer e-readers in Walmart stores and online in the U.S. starting later this year.

The content will be accessible through a Walmart-Kobo app on Apple and Android devices. Kobo offers nearly 6 million titles from thousands of publishers.

Walmart is aiming to narrow the gap between itself and Amazon online by expanding its services and adding new items. The Bentonville, Arkansas-based retailer has long offered entertainment including streaming movies through Vudu or the digital game cards it sells in its stores.  ... " 

Berkely AI Research

Newly brought to my attention:

Berkely AI Research
Overview and Blog Purpose

The BAIR Blog provides an accessible, general-audience medium for BAIR researchers to communicate research findings, perspectives on the field, and various updates. Posts are written by students, post-docs, and faculty in BAIR, and are intended to provide relevant and timely discussion of research findings and results, both to experts and the general audience. Posts on a variety of topics studied at BAIR will appear on approximately a weekly basis. .... "

Friday, January 26, 2018

Picking Your Own Produce

This has always been brought up as an issue for online grocery ordering, being unable to pick your own produce like you could in the store.   First I have heard of this approach.  I do wonder if the majority of people want that level of choice, as long as general quality was assured.

Walmart 3-D image patent lets online shoppers pick their produce
By George Anderson in Retailwire with expert discussion

(Includes patent images) Through a special arrangement, what follows is an excerpt of an article from Retail Dive, an e-newsletter and website providing a 60-second bird’s eye view of the latest retail news and trends. .... 

A new patent from Walmart would allow consumers to view real images of their groceries — rather than stock images — before purchasing online, CB Insights reported.

Through the system, called the “Fresh Online Experience” (FOE), a customer orders an item based on a stock photo. A store associate then scans the exact item in Walmart’s inventory with a 3-D scanner and the image is sent to the customer to accept or reject the particular item.  ... " 

AI in the Enterprise

AI Begins to Infiltrate the Enterprise  in InformationWeek

Enterprise adoption of AI is slow today, but experts expect it to increase very rapidly. So far, tech giants are leading the charge.

Despite a flood of publicity and product announcements related to artificial intelligence, it seems that few enterprises have adopted the technology so far.

Status Report: AI in the Enterprise

Whit Andrews, vice president and distinguished analyst at Gartner, was able to put some hard numbers to the trend. "We are in the very earliest stages of enterprise adoption of artificial intelligence," he said. "Specifically, in our most recent CIO survey from 2017, one in 25 CIOs described themselves as having artificial intelligence in action in their organizations."

The companies farthest along with the technology tend to be technology giants, said Hadley Reynolds, managing director and co-founder of the Cognitive Computing Consortium. These companies are "basing much of their businesses on various kinds of machine learning and deep learning technologies," he said, so they  .... " 

Sketching Out Your UI code

Something we always wanted to do.  At least for quick prototyping and demonstration.  But at what level of precision?   And can it easily link to existing code, data and resources?      Will take a look.  Looking forward to it.

Microsoft's 'Ink to Code' builds an app UI from your sketches
Don't throw out your napkin sketches. .... 

By Mariella Moon, @mariella_moon in Engadget .... " 

China Wants to Build AI Chips for Any Gadget

China Wants to Make the Chips That Will Add AI to Any Gadget 
In Technology Review  By Yiting Sun

China's technology industry is exploring innovative processor designs for the purpose of incorporating artificial intelligence into ordinary devices. For example, researchers at Tsinghua University in Beijing have developed Thinker, a low-power chip that is capable of dynamically tailoring its computing and memory requirements to meet the requirements of the software being run. Thinker's creators say the chip could be embedded in a broad spectrum of devices, including smartphones, watches, home robots, or equipment stationed in remote areas. The Tsinghua team is planning to launch the first product equipped with Thinker in March. Also coming soon is the production of Dadu, a dual-core chip for robots that runs neural networks and controls motion. The Dadu development team says the chip's neural core operates algorithms for vision, while the chip's motion core plans the optimal route for reaching a destination or the best motion for grabbing an object. ... " 

Thursday, January 25, 2018

Best Grocery Store in America

A place I have shopped many times.   Worth a tour.   Is telling stories a profitable grocery venture?  Can it work elsewhere?   Here is the best effort I know of.

Best Grocery Store in America
Jungle Jim's and the Art of the Tourist-Attraction Grocery Store
By Matt Siegel in Epicurious

A Los Angeleno travels to Ohio to meet the man behind the country's biggest weirdo supermarket. What he finds is a businessman who couldn't care less about food.

On the cover of the brochure for Jungle Jim’s International Market is an image of a middle-aged man in a wizard’s costume holding the notoriously stinky durian fruit high in the air as if an offering. Above the image, highlighted in yellow, it says, “FREE ADMISSION!” Free admission to a grocery store seems a given, but Jungle Jim’s is a “World of Food,” a “Foodie Wonderland,” an “International Food Lovers’ Paradise.” This according to the lengthy, hyper-color brochure, which is only a preview of the inevitably lengthy, hyper-color shopping experience that some 80,000 visitors per week enjoy at the seven-plus acre food-based freak-out in suburban Cincinnati.  ... " 

Create Common Sense Baselines

Excellent insight. Not sure that this is done frequently, even by experienced people.  We used to call this establishing base methods and estimates of results.  Sometimes just starting with:  How is it done today? or How would a domain expert do this today?  Often useful to check correctness, check likely value of results, use with domain partners to create frameworks for the use of results.  Essential to think this up front,  because if you can't, you probably don't have a full grasp of the problem.  Good read below.  With nice realistic example:

First Create a Common Sense Baseline    By Rama Ramakrishnan  Senior Vice President, Salesforce.  

When you set out to solve a data science problem, it is very tempting to dive in and start building models.

Don’t. Create a common-sense baseline first.

A common-sense baseline is how you would solve the problem if you didn’t know any data science. Assume you don’t know supervised learning, unsupervised learning, clustering, deep learning, whatever. Now ask yourself, how would I solve the problem?  .... "

Alibaba Partnering with Kroger?

Tech integrations would be interesting here.

Kroger said to discuss Alibaba partnership
E-commerce giant also operates supermarkets in China
By Mark Hamstra  in Supermarket News

Kroger Co. has held preliminary talks about a possible partnership with Chinese e-commerce giant Alibaba, according to reports on Thursday.

Executives from the two companies met last month in China, the New York Post reported. The newspaper, citing a Chinese government press release, said the two companies have already “teamed up” in an agreement “to speed up the integration of online and offline sales.”   .... " 

Anki Cozmo Robot

More personable toy robotics.  Not a new thing,  but getting us more used to the concept.

Big brain. Bigger personality.
Say hello to Cozmo, a gifted little guy with a mind of his own. He’s a real-life robot like you’ve only seen in movies, with a one-of-a-kind personality that evolves the more you hang out. He’ll nudge you to play and keep you constantly surprised. Cozmo’s your accomplice in a crazy amount of fun. ... " 

Gathering Image data in 360

A kind of life logging model,  or just image information gathering in 360 degrees

FITT360 is a 360-degree camcorder you wear around your neck
By Nicole Lee, @nicole in Engadget

Imagine if you could snap 360 degree photos and videos without holding a camera. That's the promise of the FITT360, a 360-degree camcorder that you can wear around your neck. It's one of the first products from a Samsung spin-off called LinkFlow, and was made as part of Samsung's C-Labs, the company's creative labs program.

The idea behind a wearable camcorder is to capture the world around you in an unobtrusive way. There are three cameras located on the wearable neckband -- one near the front, one on the side and another near the rear. Each camera is capable of capturing the world in 180-degrees. All you have to do is press a button and it'll start recording, allowing for a hands-free capture experience as well. And since it has WiFi, you can even use it to livestream on something like Periscope if you fancy ... "

Wednesday, January 24, 2018

Image Information for Learning in Context

Useful cognitive solutions will need to be taught, and thus need to to learn in changing contexts.  Today we re-teach most systems to adjust them to changes in environments.  The ultimate kind of augmentation.   Images in context are good examples for training material.

Vision Teacher    Technical University of Darmstadt (Germany)

Researchers at the Technical University of Darmstadt (TU Darmstadt) in Germany are teaching intelligent algorithms to detect cars, pedestrians, and potentially dangerous objects in x-ray images from transportation security. The software reconstructs image information that may be hidden by blurred or out-of-focus images, and it also could assist users in tedious tasks. To develop the system, the researchers photographed a real scene and separated the individual objects from each other by tracing their outlines. The researchers enabled the algorithms to work with less data by relying on computer games that show deceptively realistic street scenes. "Based on the information contained in the computer game, we can detect which object that is already known re-appears at a later point in time," says TU Darmstadt professor Stefan Roth. He notes this technique means the object no longer needs to be re-annotated on each video frame.   ... " 

Linear Models and Forecasting

Useful thoughts on the topic.  Being trained in this makes me think: Obviously.   But confirmation bias is easy to jump to.  Easier yet with complicated tools.  And have also seen linearity misused.

Limits of Linear Models for Forecasting
 by Blaine Bateman in DSC. .... " 

Blockchain for Retail

Had conversation on this space recently and received some eye-rolling about the hype the space has received.  Its not the same thing as the get rich quick world of 'BitCoin', just uses the same underlying tech., so deserves a closer look.   Reminds me of early reactions to the Web, claiming it was frivolous and dangerous.  Took a some time to convince management it could be an important future.  Worth reading the whole thing  and the expert comments.  There is interest and experimentation happening now.

How will blockchain disrupt retail?  by Tom Ryan in Retailwire
' .... “Blockchain can help retailers garner greater trust and brand loyalty throughout the product lifecycle, as it can tell consumers not just where an item was made, but also who it was made by, the conditions they worked in and how much they were paid,” wrote Kati Chitrakorn for Business of Fashion. It can also detect interruptions or bottlenecks so that retailers can re-route shipments or advise customers of delays.

A new IBM survey of 203 organizations in the consumer industry found only 18 percent currently using blockchain, but almost 70 percent expect to have a blockchain production network within three years. The top areas blockchain is expected to benefit: product safety and authenticity, supply chain optimization, finance/operational processes, regulatory compliance, promotional strategy management, customer engagement and co-creation. .... " 

Empirical Software Engineering

Show us the numbers.  Our of the science should come the measurable value.

Empirical Answers to Important Software Engineering Questions (Part 1 of 2)  By Bertrand Meyer 

"  ... These studies have yielded many insights including surprises. More experienced developers can produce more buggy code (Schröter, Zimmermann, Premraj, Zeller). To predict whether a module has bugs, intrinsic properties such as complexity seem to matter less than how many changes it went through (Moser, Pedrycz, Succi). Automatic analysis of user reports seems much better at identifying bugs than spotting feature requests (Panichella, Di Sorbo, Guzman, Visaggio, Canfora, Gall). More extensively tested modules tend to have more bugs (Mockus, Nagappan, Dinh-Trong). Eiffel programmers do use contracts (Estler, Furia, Nordio, Piccioni and me). Geographical distance between team members negatively affects the amount of communication in projects (Nordio, Estler, Tschannen, Ghezzi, Di Nitto and me). And so on.

The basic observation behind empirical software engineering is simple: if software products and processes are worthy of discussion, they must be worthy of quantitative discussion just like any natural artifact or human process. Usually at that point the advocacy cites Lord Kelvin:"If you cannot measure it, you cannot improve it" [5].   ... "

Tuesday, January 23, 2018

Chief Marketers View of AI

A marketing-eye view of AI.   Again taking a very broad and limited definition of the term.   Some of the ways are clearly deeply embedded in the tech we use today, so are they really new intelligence?  Or rather augmentation?  I think this will be replaced by more autonomous solutions.

 28 Ways Artificial Intelligence Will Affect You
 by Patty Odellin in Chief Marketer .... "

Beijing gets an AI District

In TechnologyReview:

Beijing Is Getting a $2.1 Billion AI District:
China is gearing up to build a technology park in Beijing entirely dedicated to the development of artificial intelligence, news first reported by Xinhua, the country’s official press agency.

Master plan: The endeavor is just the latest sign of China’s remarkable ambition to master and dominate artificial intelligence by 2020. Last year the central government published a three-year-plan to invest huge sums in AI, and to apply the technology across the country’s industries and economy (see "China's AI Awakening"). .... " 

Integrating Agent Models with Process Models

In a conversation on modeling today the topic of 'Agent Based Modeling (ABM)' came up.     And it had come to mind that such ABM models could also be usefully integrated with process models.  The ABM modeling the architecture of useful agent interaction, suitable for simulation, and a Process model simulating the sequential march towards a goal.   Has this integration been considered by anyone as a useful approach?    See past writing here on Agent models, based on work in 2012, which we used successfully to understand consumer and group behavior in retail.   We included process goals, but not formally, nor predictively in the model.

Learning from Failure

Think also, the After Action Review (AAR):

How Coca-Cola, Netflix, and Amazon Learn from Failure  By Bill Taylor

Why, all of a sudden, are so many successful business leaders urging their companies and colleagues to make more mistakes and embrace more failures?

In May, right after he became CEO of Coca-Cola Co., James Quincey called upon rank-and-file managers to get beyond the fear of failure that had dogged the company since the “New Coke” fiasco of so many years ago. “If we’re not making mistakes,” he insisted, “we’re not trying hard enough.”

In June, even as his company was enjoying unparalleled success with its subscribers, Netflix CEO Reed Hastings worried that his fabulously valuable streaming service had too many hit shows and was canceling too few new shows. “Our hit ratio is too high right now,” he told a technology conference. “We have to take more risk…to try more crazy things…we should have a higher cancel rate overall.”

Even Amazon CEO Jeff Bezos, arguably the most successful entrepreneur in the world, makes the case as directly as he can that his company’s growth and innovation is built on its failures. “If you’re going to take bold bets, they’re going to be experiments,” he explained shortly after Amazon bought Whole Foods. “And if they’re experiments, you don’t know ahead of time if they’re going to work. Experiments are by their very nature prone to failure. But a few big successes compensate for dozens and dozens of things that didn’t work.”

The message from these CEOs is as easy to understand as it is hard for most of us to put into practice. I can’t tell you how many business leaders I meet, how many organizations I visit, that espouse the virtues of innovation and creativity. Yet so many of these same leaders and organizations live in fear of mistakes, missteps, and disappointments — which is why they have so little innovation and creativity. If you’re not prepared to fail, you’re not prepared to learn. And unless people and organizations manage to keep learning as fast as the world is changing, they’ll never keep growing and evolving. .... " 

Spoofing Voice to Text

Caution for towards a need for better Assistant security.

AI learns how to fool speech-to-text. That’s bad news for voice assistants  by  Tristan Greene in Nextweb

A pair of computer scientists at the University of California, Berkeley developed an AI-based attack that targets speech-to-text systems. With their method, no matter what an audio file sounds like, the text output will be whatever the attacker wants it to be. ... "  

Actual Technical Paper.

Monday, January 22, 2018

Drone Swarms Don't Need a Break

Powerful, yet also scary scenario.    At least now they run out of power.  Note the solution here is in the process of using a swarm this way by collaborating.   The realm of our original research.

Energy Neutral Drone Swarms Can Spy on You Without Taking a Break   in Motherboard by Michael Byrne

Researchers at the University of Cambridge in the U.K. and Koc University in Turkey propose an Energy Neutral Internet of Drones (enIoD) as a swarm of aerial surveillance robots that operates continuously because charging and data transfer tasks are integrated nearly seamlessly into its function. They note the enIoD would use a hybrid energy harvesting system that does not tap the power grid, with each drone equipped with a solar cell or wind turbine, while the drones themselves serve as the grid connecting each charging station. The researchers also say wireless power transfer systems combined with wireless charging docks distributed across a wide area would enable the drone swarm to fly more or less without interruption. The team says the purpose of this network is to keep track of amateur drones, which they see as a vector for terrorists and other malefactors to attack civilians.  .... " 

AI and Auto Companies

Further, Considering automotive as a channel for goods and services

Artificial intelligence as auto companies’ new engine of value
What opportunities does AI open up for mobility, and how can OEMs capture them in the short and long run?    From McKinsey Insights   By Matthias Kässer, Asutosh Padhi, Andreas Tschiesner, and Dominik Wee ... 

Virtual Dash Buttons

Recently sent to me, an announcement that Amazon dash buttons were going 'virtual', and have been for some time.  And I was sent virtual buttons for a number of things that I had ordered in the past.  Not quite the same thing as physical buttons that make their presence on the frig, but if you reside on the Web, as many workers do, its quick and easy. 

And how is this really changing my retail behavior, when they are not physical buttons?  May depend on how organized I am online.   Would like to have  a reminder too of the relative deal I am getting.  Is it just convenience?

Dear Amazon Customer,

You can now use Dash Buttons online to instantly reorder your favorite products, just like you do with the Dash Button devices in your home. Our new virtual Dash Buttons are available for tens of millions of products that ship with Prime, and they’re free, so you can add as many as you want. We’ve automatically added buttons for products you’ve purchased in the past. Check them out today.  ....   further described here.

Cheaper Chrome Devices

Was reminded once again of the $100 laptop push of a decade ago that was aimed at schools.  Chrome Laptops are a natural next step, with much of the function needed for productivity online.   And, of course,  Microsoft plans to add cheaper mixed reality through Hololens.

Windows 10 laptops for schools. And a big education push for 2018
By Tom Warren @tomwarren in TheVerge ... "

Sunday, January 21, 2018

Supervised, Unsupervised and Reinforcement Learning

Nicely done piece, relatively non technical,  with comparison of approaches and visualizations.

Machine Learning Explained: Understanding Supervised, Unsupervised, and Reinforcement Learning - Data Science Central by Ronald Van Loon  ... 

Amazon Go Opening to Public

Don't plan to get to this anytime soon, but continue to follow automatic sensing and payment systems in detail here.   Would like to better understand the behavioral details of the interaction.  An outcome of our lab's experiments with the idea.

Amazon’s Checkout-Free Grocery Store Is Opening to the Public
At Amazon Go, you grab your milk and leave. It might take some getting used to.   by Rachel Metz  in Technology Review,
In the shadow of Amazon’s offices in downtown Seattle, people enter a tiny grocery store, take whatever they want, and then walk out. And nobody runs after them screaming.

This is what it’s like to shop at Amazon Go, the online retail giant’s vision for the future of brick-and-mortar stores. There are no checkout clerks, or even checkout stands. Instead, a smartphone app, hundreds of regular and infrared cameras on the ceiling (black on black, so they blend in), computer-vision algorithms, and machine learning work together to figure out what you’re picking up and charge you for it on a credit card connected to your Amazon account.  ... " 

Fake News Detection with AI?

We connected with Enterra a few years ago about supply chain AI. Saw their ideas presented by Stephen DeAngelis , which were impressive for the application .  See my tags on Enterra.   Now some of their work on 'Fake News' detection in Forbes:   Potential for misuse is potentially high, but worth the examination.  Also, consider how these same methods might be used to automatically detect the validity of advertising claims.  Would that radically change marketing and advertising?   Lead to an arms war between claims and detection?  No technical detail in the article.  Looking deeper.

Can Enterra's Advanced AI Systems Stop The Fake News Epidemic?
By Alan Wolk , Contributor Forbes  .... " 

And further:

 " ... For DeAngelis, the benefit to Hollywood will be the ability to apply sophisticated data-driven insights to augment the creative process. Enterra’s Enterprise Cognitive System could take a vast number of variable inputs—production schedules, star power, budgets, cultural trends, media consumption rates, and box office grosses, as well as advertising campaigns, audience targeting, and political climate—and interpret them in a way that allows for more informed executive decision making and investment.  “The ability of an AI system to think and to reason while ingesting and analyzing data from all these sources—that’s going to prove to be very powerful for media companies,” he notes.  ... " 

(Update) More on this on Enterra site.

Saturday, January 20, 2018

Challenges for AI Redesigning Healthcare

Six Challenges To Tackle Before Artificial Intelligence Redesigns Healthcare in Medical Futurist

The potential of artificial intelligence for making healthcare better is indisputable. The question is how to integrate it successfully into our healthcare systems. For doing so, we have to overcome technical, medical limitations, as well as regulatory obstacles, soothe ethical concerns and mitigate the tendency to oversell the technology. The very first step should be to familiarize with what A.I. truly means. In our latest e-book, A Guide to Artificial Intelligence in Healthcare, we summarized everything you need to know, so now, let’s look at the second step: the challenges to overcome! 

The transformative power of A.I. in healthcare

We have written extensively about the potential of artificial intelligence for redesigning healthcare. How it could help medical professionals in designing treatment plans and finding the best-suited methods for every patient. How it could assist repetitive, monotonous tasks, so physicians and nurses can concentrate on their actual jobs instead of e.g. fighting the tread-wheel of bureaucracy. By what means A.I. could prioritize e-mails in doctors’ inboxes or keep them up-to-date with the help of finding the latest and most relevant scientific studies in seconds. How its transformative power makes it as important as the stethoscope, the symbol of modern medicine, which appeared in the 19th century.

There are already great examples for its use in several hospitals: Google DeepMind launched a partnership with the UK’s National Health Service to improve the process of delivering care with digital solutions. In June 2017,  DeepMind expanded its services – first of all, its data management app, Streams, to another UK hospital. IBM Watson is used at the Alder Hey Children’s Hospital as part of a science and technology facilities council project being run by the Hartree Centre. We asked the first experiences of physicians with the technology and gave an overview of the ever-expanding line of companies who are extensively investing in the development of the technology recognizing its transformative capability in healthcare.

However, the question we always have to face is how we translate the vast potential of artificial intelligence into everyday life. After the very first step – getting to know the most possible about A.I. in healthcare -, we should get a clearer picture of the obstacles.

Get the e-book here to read about the best companies, the way A.I. will redesign healthcare, the practical examples and ethical challenges! .... "

Visualizing Uncertainty

Much of this is well known in the data visualization space, but like to see it in one place.  Should be taught this way.  Uncertainty should always be included in early analysis of data.   Includes pros and Cons of approaches.

Visualizing the Uncertainty in Data  by Nathan Yau

Lifeguards, Dogs and Light

About models and the need to create transparency through understandable models.    With yet  another tip of the hat to Phyicist Richard Feynman.   To Save Drowning People, Ask what Light Would do.  In Nautilus.

Robomart Brings an Autonomous Store to You

Unlikely I think except for narrow markets.  Interesting follow though.

Will retailers go on the road with self-driving mobile stores?   by Matthew Stern in Retailwire with expert comments.

While legislators in some cities have been trying to keep autonomous robots off sidewalks, a new startup wants to enable any big box retailer to get a full wing of its operation rolling around on city streets.

Santa Clara, CA-based startup Robomart purports to be in talks with big box retailers and wholesalers to implement its mobile autonomous grocer platform, according to TechCrunch. The company plans to allow partner retailers to use its fully-autonomous, full-sized delivery vehicles to serve as auxiliary, mobile storefronts. ... " 

Friday, January 19, 2018

New Tech Altering Business Organization

Interesting claim, though don't quite agree yet.   The tech is still valuable procedural and operational augmentation,  and not structural change.   Can see it leading to that though to drive more value.

AI, VR And AR Are Transforming Business Organisation
New research from CompTIA suggests that AI, blockchain, VR and AR are instrumental in the modernisation of global businesses.  By Kevin Joyce   in VRFocus

Amid a wave of hype around emerging technologies, three specific trends are showing the potential to help businesses transform the way they operate, according to a series of research briefs published by CompTIA. To varying degrees, artificial intelligence (AI), blockchain, virtual reality (VR) and augmented reality (AR) are becoming a more prominent technology in the digital operations of organisations, the publications reveal.   .... " 

Baidu China's Alexa

More on the elusive indications of what Baidu is up to.

China Has Its Own Alexa, But It’s No Mere Knockoff  in Fastcompany
Unlike Amazon and Google, Baidu’s DuerOS avoids competing with its hardware partners.

Baidu’s Raven H smart speaker is meant to be noticed. Whereas Amazon and Google chose to dress their speakers in neutral tones and fabric materials, the Raven H resembles a stack of green, blue, red, and orange blocks, with a partially coiled power cord that would look at home on a rotary phone. .... " 

IBM and Salesforce

For AI Style applications you must connect with experts in the domain area.    Here is an example of doing that,

IBM expands AI partnership with Salesforce    By Mike Wheatley
 IBM Corp. and Salesforce.com Inc. are expanding their partnership on all things artificial intelligence by integrating Big Blue’s Cloud and Watson services with the Salesforce’s Quip and Service Cloud Einstein products.

The company’s alliance was struck last March, when the pair announced they’d be integrating Watson with Salesforce’s relatively newer Einstein AI tech. At the time, analysts said the partnership effectively meant IBM was becoming a kind of AI consulting partner for Salesforce, allowing it to sell new services across both Einstein and Watson. .... " 

Intel RealSense 3D Camera

Had heard of Intel's Realsense project some time ago, an unusual hardware play, when they were proposed for retail signage application.  Now the devices are out and deserve a look:

Intel's new cameras add human-like 3D vision to any machine
They're ready-to-use devices meant for makers and educators.
By Mariella Moon, @mariella_moon in Engadget .... "

Thursday, January 18, 2018

Journies in Wikipedia

Here I was less interested in the fact that people spent large amounts of time linking around the Wikipedia,  I knew that,  than the particular journies they took.  Now if only they accepted advertising.

Wikipedia explains how those late-night reading binges happen
Most people visit another link when they look up a topic on Wikipedia. .... " 

By Mariella Moon, @mariella_moon in Engadget

Expanding the Dash Button

During our early work on an 'ordering button' we imagined them integrated into consumer products, right next to the point of need.  Most of my own Dash Buttons are on the frig after early concept and operations testing.   These can also detect patterns in ordering that can drive other replenishment behaviors. I would also like easier price comparison.  A Moment of truth that combines use and selection?

Amazon ramps up Dash Button integration for super-quick ordering     By Trevor Mogg in Digitaltrend 

" ... Amazon already has a number of manufacturers working the technology directly into their products, and this week it announced that more makers are getting on board.

New devices offering automatic ordering via Dash include printers from HP and Epson so you never run out of ink, a range of home appliances from Kenmore, smart air filters from 3M that automatically reorder replacement filters, coffee machines from illy that track capsule usage, and Wi-Fi “pet camera” maker Petcube that’ll automatically reorder animal treats.

Amazon this month also announced the Virtual Dash Button Service (VDBS) that allows third parties to offer virtual Dash Buttons on their screened devices — think smart washers and refrigerators.

“Virtual Dash Buttons are shortcuts that allow Prime members to quickly find and reorder their favorite products from a selection of millions of eligible products,” Amazon explains. “Virtual Dash Buttons launched on Amazon’s mobile application and website in January 2017, expanded to Echo Show in October, and now can be available on third party devices via VDBS.”

The system enables you to quickly bring up multiple items on the display — so for your washer it might be different brands of detergent and fabric conditioner — and order in a couple of taps. While the ordering is manual rather than the automatic system employed by many Dash devices, this method allows you to select from multiple products associated with a particular machine. .... " 

Google ML Automation Service

We will continue to see more automation in driving machine learning problems.   Here the training aspects for some kinds of problems are addressed, like image recognition. 

Cloud AutoML Alpha   by Google
Train high quality custom machine learning models with minimum effort and machine learning expertise

Train Custom Machine Learning Models

Cloud AutoML is a suite of Machine Learning products that enables developers with limited machine learning expertise to train high quality models by leveraging Google’s state of the art transfer learning, and Neural Architecture Search technology.

AutoML Vision is the first product to be released. It is a simple, secure and flexible ML service that lets you train custom vision models for your own use cases. Soon, Cloud AutoML will release other services for all other major fields of AI.   ....  "

Google’s AutoML lets you train custom machine learning models without having to code   by Frederic Lardinois (@fredericl)  in TechCrunch

Google today announced the alpha launch of AutoML Vision, a new service that helps developers — including those with no machine learning (ML) expertise — build custom image recognition models. While Google plans to expand this custom ML model builder under the AutoML brand to other areas, the service for now only supports computer vision models, but you can expect the company to launch similar versions of AutoML for all the standard ML building blocks in its repertoire (think speech, translation, video, natural language recognition, etc.).  .... " 

Hospitality Robotics

Clear that robotics will increasingly do labor intensive tasks with robot helpers.   Flexibility will be needed.   Indoor location and navigation solutions will be needed.

The Relay hotel delivery robot will soon spot Wi-Fi dead zones and mingle with guests   By Paul Miller  @futurepaul in TheVerge

 Directly adjacent to the Las Vegas Convention Center is a Renaissance hotel with a pair of special staff members: robots. Savioke’s Relay robots have been on the job for three months, helping out the concierge by delivering items to guests during peak hours. The two robots, named “Elvis” and “Priscilla” by the hotel, pick up orders from the front desk, call and ride the elevators without help, and call the guest’s room phone when they’ve arrived. They navigate autonomously, based on a pre-generated map, so there’s no problem if they lose Wi-Fi or LTE signal. I got to watch a delivery in action (to a demo room), and it was a seamless experience. ,,,, " 

Wednesday, January 17, 2018

Kodak and Blockchain to Control Photos

We visited Kodak innovation just before the digital revolution completely flattened and changed the film business.  So their history of innovation is interesting to me. Also later worked on the business of monetizing photographs, so this particular application is instructive.   This could also be a way to demonstrate and familiarize the idea of Blockchain to the consumer.

Kodak’s stock doubles after camera and film pioneer boards blockchain bandwagon  By Emily Bary in Marketwatch

CEO says blockchain is more than a ‘buzzword’ and will help photographers assert control over their work  ... "

Kroger to Expand Kroger Edge

We studied many kinds if shelf based technologies, below seem to be the outcome of many of these ideas tested.    It is mentioned in the article that it was developed with Microsoft technologies like Azure, and  that Kroger plans to sell it to other retailers.

Kroger is rolling out a new technology to nearly 200 stores that could change grocery shopping as we know it  By Hayley Peterson in Business Insider

Kroger is rolling out a new technology that will communicate with customers' smartphones and highlight products on their shopping lists. Kroger

Kroger is rolling out a new technology, called Kroger Edge, to nearly 200 stores by the end of 2018, the company told Business Insider.

The technology digitally displays pricing and nutritional information, as well as video ads and coupons for various products.

It will eventually communicate with customers' smartphones and highlight products on their shopping lists as they walk down store aisles — helping them to quickly select items in a sea of identical-looking items.

It will also highlight items that match customers' dietary needs if they have allergies or other dietary restrictions.   .... " 

Windows Insider on Mixed Reality

Been following Microsoft's interest in 'mixed reality' aka augmented reality for a client.  Just discovered Microsoft insider podcasts, which looks at near future directions in Windows operating systems,  applications and their potential use in the enterprise.  Most recently they have just featured a considerable podcast on mixed reality that is worth a look.  Quote:

Alex Klipman:  (from video):  Now we're standing together at the threshold of the next revolution of computing.  Now, the thing that excites me about this revolution is that computers will empower us to renegotiate our very contract with reality, giving us the capability to transcend time, space, and devices

In this revolution we will immerse ourselves in virtual worlds of our choosing, and we'll be able to accomplish impossible things.  And we'll be able to do all of this while creating lasting memories with the people that we love

Our very sense of reality is set to be transformed as we enter this new era of computing, the era of mixed reality .... " 

Stewart Brand Speaks at Long Now

I was introduced to the Long Now Foundation soon after its inception, and used to mention its efforts often in the predecessor of this blog.  Also Stewart Brand, founder of the Whole Earth Catalog, and The Welloften met with us at the Institute for the Future.  Not much in recent years, but just recently:

Stewart Brand Gives In-Depth and Personal Interview to Tim Ferriss

Tim Ferriss, who wrote the The Four Hour Work Week and gave a Long Now talk on accelerated learning in 02011, recently interviewed Long Now co-founder Stewart Brand on his podcast, “The Tim Ferriss Show”. The interview is wide-ranging, in-depth, and among the most personal Brand has given to date. Over the course of nearly three hours, Brand touches on everything from the Whole Earth Catalog, why he gave up skydiving, how he deals with depression, his early experiences with psychedelics, the influence of Marshall McLuhan and Buckminster Fuller on his thinking, his recent CrossFit regimen, and the ongoing debate between artificial intelligence and intelligence augmentation. He also discusses the ideas and projects of The Long Now Foundation. ... "

Roomba Beta Advances

iRobot Launches Beta Program to Test Smarter Home Features for Roombas  By Evan Ackerman  in IEEE Spectrum

Robots like the 900 series Roomba, which have the ability to make detailed maps of your home while they do chores, are poised to become much more useful than they've ever been. These maps are in the process of enabling all kinds of new features, most of which we can only guess at, but iRobot will be giving us a peek at what they're working on with a late Christmas present: a new beta program for Roombas.

In a next step to make the smart home smarter, iRobot today announced a new Beta feature that allows customers to view a Wi-Fi Coverage Map created by Roomba. The Wi-Fi Coverage Map feature is the first to be available as part of iRobot Beta, a program that enables a limited number of iRobot customers to trial unreleased product features for the company’s consumer robots. iRobot Beta is accessed through the iRobot HOME App and will be available in mid-January (US only).

WiFi Map
Wi-Fi Coverage Mapping is a Beta feature that enables Roomba 900 Series vacuums to detect WiFi signal strength throughout a user’s home to highlight areas that may have poor connectivity. Users can then work to determine ways to fix Wi-Fi issues, resulting in an improved customer experience when using connected devices within the home. Users simply start a Roomba cleaning job, and the robot will collect Wi-Fi signal information as it cleans. Upon completion, users will receive a notification to view the Wi-Fi Coverage Map that was generated. The user will be able to toggle between the familiar Clean Map™ Report, which shows where Roomba cleaned, and the Wi-Fi Coverage Map. ... "

Tuesday, January 16, 2018

Drone Swarms Flying without GPS

Brought up in a recent challenge for Drone Swarms and Drones indoors.

This Autonomous Quadrotor Swarm Doesn't Need GPS    By Evan Ackerman in IEEE Spectrum

UPenn's autonomous quadrotor swarm doesn't need GPS or external localization to fly indoors or outdoors. .... "

CES: More Assistants in the Car

CES: Amazon's Alexa Making More Inroads Into Vehicles
David Kiley , Contributor.   Providing insights and news about the global auto business.

For all the talk about Google and Apple getting into the auto business, it is tech giant Amazon that is making headway getting into the car business as more car companies integrate the online retailer’s Alexa voice-activated personal assistant into their in-car connected services.

The latest automaker to join forces with Amazon is Toyota, which will start adding Alexa connectivity to some of its vehicles starting this year. Alexa will enable drivers and passengers to get directions, control in-car infotainment services and perform all the other ask-and-answer functions Alexa delivers in the home application. ... "

Maersk, IBM Launch Blockchain

Good to think about the value proposition here,  how it is beyond a simple ledger database approach.  The stated value is to have controlled data sharing among those shippers and other parties that sign up for the service.  Back to the idea of verified transparency which should lead to better data and thus analytics.  No mention of Smart Contracts, but seems to be a logical next step.

Maersk, IBM to launch blockchain-based platform for global trade  By Jacob Gronholt-Pedersen in Reuters

COPENHAGEN (Reuters) - The world’s largest container shipping firm A.P. Moller-Maersk is teaming up with IBM to create an industry-wide trading platform it says can speed up trade and save billions of dollars. ... " 

In Home Delivery Expands

There seemed to be quite a bit of push back about giving retailers the ability to open your home door and drop a package inside.  Some horror stories emerged.  But the idea seems to be continuing to grow. With more details and Expert discussion:

Smart home partnership opens doors to more in-home delivery  by Matthee Stern in Retailwire

" ... August Homes, a smart home locks startup, which partnered with Walmart on its initial in-home delivery initiative, is now working to enable the service for customers of delivery startup Deliv, according to CNBC. Deliv’s retail accounts include Macy’s, Plated, Kohl’s, Best Buy and Walgreens. As with the Walmart service, a Deliv employee who is making a smart home delivery will be given a unique code that gives them one-time access to the residence. August Homes is planning relationships with other delivery services as well. .....  " 

Alibaba Claims to Beat Humans in Reading Test

More examples of advanced AI in China.

Alibaba's AI Outguns Humans in Reading Test    By Robert Fenner in Bloomberg

Alibaba has developed an artificial intelligence model that scored better than humans in a Stanford University reading and comprehension test.

Alibaba Group Holding Ltd. put its deep neural network model through its paces last week, asking the AI to provide exact answers to more than 100,000 questions comprising a quiz that’s considered one of the world’s most authoritative machine-reading gauges. The model developed by Alibaba’s Institute of Data Science of Technologies scored 82.44, edging past the 82.304 that rival humans achieved.

Alibaba said it’s the first time a machine has out-done a real person in such a contest. .... " 

Neural Network Zoo Updated

Kirk Bourne writes that the useful  'Neural Network Zoo' now contains links to the original papers for each network type AND a poster version of the full infographic.

In the Asimov Institute, post by Fjodor Van Veen

" .... With new neural network architectures popping up every now and then, it’s hard to keep track of them all. Knowing all the abbreviations being thrown around (DCIGN, BiLSTM, DCGAN, anyone?) can be a bit overwhelming at first.

So I decided to compose a cheat sheet containing many of those architectures. Most of these are neural networks, some are completely different beasts. Though all of these architectures are presented as novel and unique, when I drew the node structures… their underlying relations started to make more sense.  ... " 

Imposter Syndrome and the Turing Test

How a Famous Robot Test Can Help You Beat Impostor Syndrome  by Kelton Reid in Copyblogger

Points to a 2012 Atlantic article on using the Turing Test to discern (or fool) people, or even ourselves into determining who we are.   Fascinating piece,  especially with regard to how an intelligent dialog is considered part of the test process.  Which is different than just asking a question and determining if people can answer it correctly.  Its not just what you know, but how you can adjust that knowledge to contextual needs.    Also harder for AI to reproduce.

" ... In the words of sociologists, what they're now studying is called "interactional expertise." The easiest way to understand what interactional expertise entails is to contrast it with a more common idea, contributory expertise. Contributory experts are the typical array of professionals (physicists, chemists, lawyers, economists, musicians etc.) who develop specialized knowledge and skill through formal education and long experience. ... "  

Monday, January 15, 2018

Crucial Need for Uncertainty in Deep Learning

Below piece really rocked me.  It is very old AI school.  Even takes us to methods that were used in  Watson-Jeopardy.  Conclusions, and even sub conclusions, are rarely precise answers.  They need to contain a certainty factor (CF).   All the old AI systems embedded certainties.   Trouble is, will the methods still converge the way the new methods do?

Our own real-problem based research showed that was not always certain itself.  The need came up in recent modeling work.  Thus the Google research.  Glad to see this is being brought out, but it may slow movement in AI.   This is a huge thing,  but uncertainty must be considered.

Google and Others Are Building AI Systems That Doubt Themselves
AI will make better decisions by embracing uncertainty.   by Will Knight  in Technology Review

The most powerful approach in AI, deep learning, is gaining a new capability: a sense of uncertainty.

Researchers at Uber and Google are working on modifications to the two most popular deep-learning frameworks that will enable them to handle probability. This will provide a way for the smartest AI programs to measure their confidence in a prediction or a decision—essentially, to know when they should doubt themselves.

Deep learning, which involves feeding example data to a large and powerful neural network, has been an enormous success over the past few years, enabling machines to recognize objects in images or transcribe speech almost perfectly. But it requires lots of training data and computing power, and it can be surprisingly brittle.

Somewhat counterintuitively, this self-doubt offers one fix. The new approach could be useful in critical scenarios involving self-driving cars and other autonomous machines.  .... "