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

Tuesday, March 20, 2018

IBM Delivers a Watson Voice Powered Assistant

Been looking at this in Beta for some time.  More detail to follow.

IBM delivers Watson-powered voice assistant for consumer brands
Alexa and Google Assistant have taken residence in people's homes. IBM aims to give companies a way to deliver their own branded AI voice assistants

IBM has launched Watson Assistant, an artificial intelligence (AI) powered voice assistant for businesses.

Organisations showcasing the Watson Assistant include speaker maker Harman, retail bank Royal Bank of Scotland, Autodesk, Munich Airport and Motel One.  ... " 

Behavioral Implications of Grab and Go Retailing

Some interesting behavioral observations of early use of the lack of checkouts in Amazon's Grab and Go tests.   We interviewed and watched consumers in our laboratory stores to learn how they felt and reacted to similar approaches.  Will this cause fewer visits, change the nature of visits and purchases?  How will it interact with online visits?    Will this be an ultimate expectation of physical stores?   Amazon is in position to learn much here.

Amazon Go customers are still adjusting to the grab-and-go model in DigitalTrends

Apparently, our parents have taught us well. While Amazon’s new cashless grocery store, Amazon Go, has encouraged folks to just walk out the door without paying, it would seem that folks aren’t quite on board with that model yet. According to Gianna Puerini, vice president of Amazon Go, it has taken shoppers a bit of time to get used to the fact that walking out of a store without stopping by a cash register is not, in fact, immoral or illegal.

At Shoptalk, a retail industry event in Las Vegas, Puerini noted that she has been struck by the number of customers who have second-guessed their ability to take advantage of the cashless convenience offered by Amazon Go. ‘‘What we didn’t necessarily expect was how many people would stop at the end on their first trip or two and ask, ‘Is it really OK if I just leave?’’’ Puerini said of the new-age store that opened in January in Amazon’s hometown of Seattle. .... " 

Monday, March 19, 2018

Optimizing Health Policies with Bayesian Networks

 Another excellent, mostly nontechnical presentation on the topic.   Interesting is the decision model itself, and the topic of health decisions.  Unlike most modeling methods, this approach embeds the details of the model into the decision process being modeled.  So you can visually see the details of what is being modeled and discuss it with decision makers.  Also, it directly models uncertainty involved, based on real known data.   We used these methods actively,  I but find them rarely applied in business.   Consider it.  ....

Presentation link and slides below: 

By Stefan Conrady

Managing Partner at Bayesia USA & Singapore: Bayesian Networks for Research, Analytics, and Reasoning

Optimizing Health Policies with Bayesian Networks

In case you missed today's webinar, here is the recording. Today's program was about developing a reasoning framework for health policies in developing nations with Bayesian networks. The specific study question was whether to implement a "test & treat" policy versus a presumptive treatment approach for malaria and bacterial pneumonia. https://bayesia.wistia.com/medias/16vb2vljlt

HBR: Getting Value from Machine Learning

Makes a very obvious case.    That has existed since the beginning of computing.  Yet still a good one to repeat.  Systems must be easy enough to use.  And then actually used, to make them valuable.  Of course when you add in some level of autonomy, with a clearly measurable value, that helps.   One way to do that is to plug them into a known and measurable business process.   You can show the value of it being better, faster, or cheaper, directly.  Augmentation of people and processes is best.  Making the method automatically considered and even applied.  We did it many times.  Good example of project management below. 

Getting Value from Machine Learning Isn’t About Fancier Algorithms — It’s About Making It Easier to Use    By Ben Schreck, Max Kanter, Kalyan Veeramachaneni, Sanjeev Vohra, Rajendra Prasad  in the HBR

Machine learning can drive tangible business value for a wide range of industries — but only if it is actually put to use. Despite the many machine learning discoveries being made by academics, new research papers showing what is possible, and an increasing amount of data available, companies are struggling to deploy machine learning to solve real business problems. In short, the gap for most companies isn’t that machine learning doesn’t work, but that they struggle to actually use it.

How can companies close this execution gap? In a recent project we illustrated the principles of how to do it. We used machine learning to augment the power of seasoned professionals — in this case, project managers — by allowing them to make data-driven business decisions well in advance. And in doing so, we demonstrated that getting value from machine learning is less about cutting-edge models, and more about making deployment easier.  .... " 

Inferring Emotion and Cognitive Changes

The OBAIS department at the Lindner College of Business, University of Cincinnati, invites you to attend a research seminar:

Date and time: Wednesday, March 28th, 2018, 11:00AM-12:00PM

Location: Lindner Hall 608
Speaker: Prof. Joe Valacich, Eller Professor in MIS, University of Arizona

Title: Inferring Emotion and Cognitive Changes through Human-Computer Interaction Devices: From Basic Research to Communalization

Best wishes,

Yichen Qin, Assistant Professor
Department of Operations, Business Analytics, and Information Systems
Lindner College of Business, University of Cincinnati
Website: http://business.uc.edu/academics/departments/obais/faculty/qinyn.html
Email: qinyn@ucmail.uc.edu

Measuring Results

How Accurate Is Your AI? 
from Kyoto University

A researcher at Kyoto University in Japan has developed a new technique that evaluates artificial intelligence's (AI) performance based solely on the input data. In typical AI development, a performance evaluation is trusted if there is an equal number of positive and negative results, and data biased toward either value means the current system of evaluation will distort the system's ability. "The novelty of this technique is that it doesn't depend on any one type of AI technology, such as deep learning," says Kyoto's J.B. Brown. "It can help develop new evaluation metrics by looking at how a metric interplays with the balance in predicted data. We can then tell if the resulting metrics could be biased." Brown's work breaks down the AI utilization and analyzes the nature of the statistics used for reporting an AI's ability, while also producing a probability of the performance level, given evaluation data. .... " 

Testing and Automation of Assistant Skills

Like paying attention to the process of creating and delivering skills: 

Building Engaging Alexa Skills: Why Testing and Automation Matter

By Paul Cutsinger  In Amazon Developer

By Editor’s Note: Skill testing is one of the most important things you can do to build high-quality voice experiences. Today we welcome a community expert in testing tools for voice—John Kelvie, founder and CEO of Bespoken—to share some best practices.

Developing for Alexa can be a lot of fun. There are so many opportunities to create innovative user experiences. The cutting edge is constantly evolving. And the reachable audience is immense, and always expanding.

When building skills, it is incredibly important to build high-quality experiences for users. These users will not come back if a skill does not open or fails quietly halfway through. And we may not be aware of any problems until a user writes a one-star review. This is not the ideal way to identify and fix bugs; there must be a better one.

And there is. Testing and automation are the solution. They help us deliver reliable skills for customers and a great user experience. Through testing and automation, we can offer consistently great experiences to our users. This blog will outline how to do this at a high level and also offer some practical steps to implement it. .... " 

Online Grocery to Reach $100 Billion

Online grocery sales could reach $100 billion by 2022, researchers say    By Andrea Miller   ABC

Walmart plans on expanding grocery delivery to 100 metropolitan areas

Instead of taking a trip to a local grocery store, more and more consumers are opting to order essentials such as cereal, toothpaste or even apples from online retail giants.

Walmart, for one, announced Wednesday that it is bolstering its grocery-delivery service to reach even more cities. .... " 

Macy's Using Virtual Reality for Furniture Sales

Really a pretty old idea, was one of the first ideas we examined for demonstration and sales.    I encountered IKEAs approach in-store  just a few days ago, well done, but not enough AR to understand how your choices would fit in.   Also drove home the point that for store and online experiences the consumer needs to be able to use the system quickly.  Its different in research.   We experimented with it to understand how product would exist on shelves with other products.   See also approaches that mix VR, AR and physical digital displays, such as John Milby's Full Scale Virtual Research (FSVR).

Macy’s will use VR to sell furniture in 50 stores by summer
 By Jeremy Horowitz  @Horowitz  in Venturebeat

VR and online shopping are often portrayed as enemies of brick-and-mortar retail, but shopping mall anchor Macy’s plans to embrace both technologies in a bid to improve its sales, reports FurnitureToday. Speaking at the ShopTalk retail conference in Las Vegas, Macy’s CEO Jeff Gennette announced that he will bring VR furniture-selling tools to 50 stores by this summer and plans to offer the immersive shopping technology in “as many stores as possible.”

According to Gennette, the virtue of virtual reality is its ability to “sell more furniture with less, or even no, square footage devoted to displaying it.” Macy’s piloted a VR system that let customers use a tablet to add furniture to a room, move the pieces around until they seemed optimal, then experience the fully furnished room using VR. The system enabled customers to feel more comfortable about furniture fit and “significantly increased” both total transaction sizes and sales of items that Macy’s carries but didn’t keep on site.  .... " 

AI for Competitive Value

Everyone is still asking,  how much of this is hype?  There is an element of that, but clearly value as well.  How much should the enterprise invest?

How machine learning is changing the game for app marketers  in Thinking with Google    ... Jason Spero Nov 2017 Apps, Emerging Technology, Mobile, Data & Measurement

Artificial intelligence and machine learning technology have the potential to revolutionize marketing as much as mobile, the internet, and television did in the past.

Forward-thinking companies are using machine learning tools to supercharge their marketing. These early adopters take advantage of the technology’s ability to streamline data, unlock user insights, and engage users in highly relevant ways. In fact, 85% of executives believe AI will allow their companies to obtain or sustain a competitive advantage, according to the The Boston Consulting Group.  ... "

Smart Speakers Addictive?

In what sense?  Because they are frequently always on, perhaps, but because they use voice as assistants I find  myself using them less than smartphone in public or semi-public situations.

How addictive are smart speakers?  by Tom Ryan  in Retailwire.

According to the latest Smart Audio Report from NPR and Edison Research, 65 percent of voice-activated smart speaker owners “wouldn’t want to go back to life without” their Amazon Echo or Google Home.

That finding exceeded the 46 percent of Americans who told Pew Research Center in 2014 they “couldn’t live without” their smartphones.

One caveat from NPR’s survey last November of 1,800 consumers is the finding that only 16 percent of Americans own a smart speaker. But the survey still demonstrated how smart speakers are changing behaviors and causing owners to form new habits.

For instance, when smart speaker owners were asked what other devices they are spending less time with as they use they increase their smart speaker usage, the top answer was traditional radio, at 39 percent. That was followed in the top-five by smartphones, 34 percent; television, 30 percent; tablets, 27 percent; and computers, 26 percent.  .... "

Sunday, March 18, 2018

Driverless Pizzas to be Delivered before People

Inclined to generally agree, general driver less delivery should precede driver-less vehicles with passengers.   If only for the liability and legal issues involved.    Yet driver-less vehicles will come.  But agree less with the article that we will soon see many customers meeting the driver-less delivery vehicles out by the curb to eliminate the last 100 yards.   It is still extreme convenience that is leading this transition.   Thoughtful piece on business process profitability issues:

Why Self-Driving Vehicles Are Going to Deliver Pizzas Before People     By Bloomberg in Forbes

In the wait for self-driving technology, cell-phone toting tech bros may have to cede their spot in line to pizzas, Craigslist couches and the mounting ephemera of e-commerce.

The future—at least in the near-term—will not only be driverless, but sans passenger as well.

The early conversations around driverless cars have focused on robot taxis because taking the human driver out of a cab seemed like the quickest path to profitability. But an increasing number of companies—automakers, tech giants, startups, parcel services—are seeing autonomous delivery as the more lucrative venture.

“The revolution in commercial vehicles will come first, then the passenger cars” will follow, said Ashwani Gupta, senior vice president of Renault-Nissan’s light commercial vehicle business. “The moment business people start believing this is going to generate additional revenue and that this is going to be more efficient, then I think they’ll start working on it.”  ... ' 

Fujitsu Human Centric AI

Was impressed with Fujitsu's work in retail when we visited.

Fujitsu drives a human centric model

AI is a core technology which enables many complex processes to be conducted independently of human judgment. Now, deep learning is often featured in the media. But it is not the whole story of AI, just an important piece of the puzzle. Our human cognition is continuously generated from complex interactions between our sensory organs, nervous system, brain and external environments.

To achieve an AI, we have to replicate and bring together a range of cognitive capabilities: perceiving, reasoning, making choices, learning, communicating, and moving and manipulating.

Fujitsu is developing key technologies under a comprehensive framework (see diagram). We call it Human Centric AI, Zinrai. Fujitsu is incorporating component technology such as machine learning, deep learning and visual recognition, into its digital solutions and services. .... " 

Baidu's AI Mimicing Voice

Baidu’s new A.I. can mimic your voice after listening to it for just one minute
By Luke Dormehl in Digital Trends

" ... “From a technical perspective, this is an important breakthrough showing that a complicated generative modeling problem, namely speech synthesis, can be adapted to new cases by efficiently learning only from a few examples,” Leo Zou, a member of Baidu’s communications team, told Digital Trends. “Previously, it would take numerous examples for a model to learn. Now, it takes a fraction of what it used to.” .... " 

Samsung TVs Controlled with Bixby Assistant

Rumor out there that perhaps Samsung would tap Alexa and/or Google for voice control.  But it seems they are sticking with their own Bixby voice assistant for controls.  Up to now only on phones, that will likely soon change.  Looking to test.

Samsung TVs tap Bixby for voice, SmartThings for home control

The company improves its already excellent Smart TV system with an app for easy setup and smart home control, as well as the Bixby voice assistant.     By David Katzmaier ...  

Misinformation and the Wikipedia

Been a longterm Wikipedia fan.   And have also been directed to many, many examples of misinformation there.  But still use it daily.  So whats the solution?  Apparently Youtube planning to resource credibility with WP articles, among others.    Further curated?   Only as good as the curators.  Bias is all over the place.

Don't ask Wikipedia to Cure the Internet   by Louise Matsakis in Wired.

" .... On stage at the South by Southwest conference on Tuesday, YouTube CEO Susan Wojcicki announced that her company would begin adding "information cues" to conspiracy theory videos, text-based links intended to provide users with better information about what they are watching. One of the sites YouTube plans to use is Wikipedia. "We’re just going to be releasing this for the first time in a couple weeks, and our goal is to start with the list of internet conspiracies listed where there is a lot of active discussion on YouTube," Wojcicki said on stage..... " 

Saturday, March 17, 2018

Tags in this Blog

This blog contains tags at the end of each post which lead to related posts.   I do go back selectively and update these tags, especially as they relate to my current research, interests or work. The tags can't be complete,  in some cases the tag topic may not exist until much later.     For example a company that is later formed to address some new technology.  This blog is for my own and client reference,  but if you have any suggestions pass then along in a comment or email.  I am on Linkedin and will respond there too.   - FAD

(Updated) Optimization using Genetic Methods

In our earliest days,  addressing supply chain and blending type manufacturing problems, we were an optimization shop.  Using the math structure of difficult combinatorial problems to find best solutions based on known goals and constraints.    But if you couldn't glean enough low level structure, we tested genetic methods, described here.   In this era of faster machines and more contextual information even more useful to try today.  Also for certain kinds of structure, also consider Dynamic Programming.  Happen to be examining that again today.

In KDNuggets  By Ahmed Gad, KDnuggets Contributor 

This article gives a brief introduction about evolutionary algorithms (EAs) and describes genetic algorithm (GA) which is one of the simplest random-based EAs.

Selection of the optimal parameters values for machine learning tasks is challenging. Some results may be bad not because the data is noisy or the used learning algorithm is weak, but due to the bad selection of the parameters values. This article gives a brief introduction about evolutionary algorithms (EAs) and describes genetic algorithm (GA) which is one of the simplest random-based EAs.


Suppose that a data scientist has an image dataset divided into a number of classes and an image classifier is to be created. After the data scientist investigated the dataset, the K-nearest neighbor (KNN) seems to be a good option. To use the KNN algorithm, there is an important parameter to use which is K. Suppose that an initial value of 3 is selected. The scientist starts the learning process of the KNN algorithm with the selected K=3. The trained model generated reached a classification accuracy of 85%. Is that percent acceptable? In another way, can we get a better classification accuracy than what we currently reached? We cannot say that 85% is the best accuracy to reach until conducting different experiments. But to do another experiment, we definitely must change something in the experiment such as changing the K value used in the KNN algorithm. We cannot definitely say 3 is the best value to use in this experiment unless trying to apply different values for K and noticing how the classification accuracy varies. The question is “how to find the best value for K that maximizes the classification performance?” This is what is called optimization.

In optimization, we start with some kind of initial values for the variables used in the experiment. Because these values may not be the best ones to use, we should change them until getting the best ones. In some cases, these values are generated by complex functions that we cannot solve manually easily. But it is very important to do optimization because a classifier may produce a bad classification accuracy not because, for example, the data is noisy or the used learning algorithm is weak but due to the bad selection of the learning parameters initial values. As a result, there are different optimization techniques suggested by operation research (OR) researchers to do such work of optimization. According to [1], optimization techniques are categorized into four main categories:  .... " 

  (Update) A comment I got made me add this.  'Optimization' in business practice implies you can get the provably, best possible solution to a problem.   But in reality it almost always means you only can get the best solution within some specific context.     A context can include structure, constraints and goals.    It may also vary over time.    It may be wrong because its too hard to completely understand the problem.  But its still often useful to get a better solution, even if not provably optimal, if its better than todays practice.     Further if you can calculate this 'theoretical' best solution, it can give you better understanding of a problem, and what to strive for.    - FAD 

Sentier Strategic Consulting

Sentier Strategic Resources conducts custom research that helps our clients better understand and respond to the behavior, perceptions, and expectations of their market.

The Sentier team is comprised of experts in cognitive science, social psychology, linguistics, data science and human factors engineering, each having PhDs and years of experience in finding answers to real world questions, both big and small. We think about your research questions as deeply as you do, using data to understand the “why” and advise you what it means for your business. We understand that when you hire a research firm, your reputation is at stake within your organization. We take that seriously and will make you look great by providing research that is accurate, usable, methodologically sound, and executive-presentation ready.

So why work with Sentier?

Expertise: Our experience in consumer and enterprise technology ecosystems enables us to design an approach that answers your specific business questions.

Dexterity: We design, manage, and tailor projects to integrate with your systems, processes, and timelines. We work in multi-national and multilingual contexts.

Impact: We deliver useful analysis with valuable recommendations, and illustrate them with clear, effective visualization, as well as photo and video assets.

Sentier was founded in 2009 and is headquartered in Austin, Texas USA. We are a woman-owned small business. We are HIPAA compliant and SAM registered.   ...

Friday, March 16, 2018

Augmented Beauty by Modiface at L'Oreal

An area we did lots of research and development on.  Now based on this piece, it seems that the tech has finally caught up to the needs.   But will it practically work as a marketing, sales and operational tool?  Remains to be seen.   See images at the link.

L’Oreal acquires Modiface, a major AR beauty company
By Ashley Carman @ashleyrcarman  in TheVerge

L’Oreal announced today that it has acquired Modiface, a company that’s had a hand in the creation of many custom augmented reality beauty apps, including those from Sephora and Estée Lauder. L’Oreal didn’t disclose the amount spent, but it did tell Reuters that it now owns Modiface’s numerous patents that help users visualize makeup and hairstyles on themselves. The partnership makes sense in that Modiface has already worked with L’Oreal multiple times, including on the launch of its Style My Hair mobile app, which lets users try on different hairstyles. For that app, Modiface manually annotated 22,000 facial images to create the experience.  ... "

Iterative Random Forests

New Learning Method:  Sees the Forest and the Trees

".... Researchers at the Department of Energy's Lawrence Berkeley National Laboratory (Berkeley Lab) and University of California, Berkeley have created a novel machine learning method that enables scientists to derive insights from highly complex systems in record time.

In a paper published recently in the Proceedings of the National Academy of Sciences, the researchers describe a technique called "iterative Random Forests," which they say could have a transformative effect on any area of science or engineering with complex systems, such as biology. ... "

Amazon Pickup Service in Whole Foods

Witnessed the set up of this in a nearby Whole Foods today.  No additional crowding as yet.   Can see it as a specialized service offering, volume involved unclear.  Other uses when the infrastructure is operating?

Amazon/Whole Foods planning store pickup service from third-party retailers  by George Anderson in Retailwire.  with further expert comments:

Amazon.com wants to negate one advantage that rivals such as Walmart, Target, Kroger and others have — store pickup. The e-tailing giant is looking to offer a pickup service at Whole Foods’ stores that will not only include orders from the organic grocery chain, but also from a host of other retailers.

According to the reports, Amazon is seeking a finance manager that will help build a pickup business from the ground up. The job posting, which was first reported on by the Puget Sound Business Journal, said the person hired would be behind “the Whole Foods delivery and pick-up service on the ultra-fast Prime Now app and enable our Prime customers to shop from a set of marquee third-party retailers.”

What potentially makes the described service different from those offered by Walmart and others is that it would appear to offer pickup from online orders placed with Whole Foods, Amazon and perhaps others, as well.  ... " 

Ring and Amazon

I am a user of the Ring Doorbell, have been since their beginning.   So intrigued by the implications. New kinds of image data collection?  Amazon Key service has been covered here.  Privacy of behavior in the home.

What does Ring mean for Amazon?   in Retailwire  by Chris Petersen with expert comments. 

Through a special arrangement, presented here for discussion is a summary of a current article from the IMS Results Count blog.

Amazon.com in late February acquired Ring, a maker of internet-connected doorbells and cameras, for about $1.1 billion.

Ring is best known for its Wi-Fi enabled doorbells that are equipped with cameras to detect when someone is at the door. Users receive an alert and then are able to view and talk to the individual outside their door through their smartphone.

On the surface, Ring is a powerful acquisition, which launches Amazon further into the home security space. Last year it began selling Amazon Cloud Cam, an indoor security camera of its own design. In December it acquired Blink, a maker of inexpensive internet security cameras and doorbells. Amazon also moves further into the IoT space with more popular products that can connect to Alexa. Google’s Nest also offers a home security system.

The apps and Ring subscriptions will create recurring revenue. All well and good in itself, but several reports on the acquisition focused on how Ring’s technology may build on Amazon Key, a service launched last October that allows Prime members to have orders delivered inside their homes to help deter theft and prevent fresh food from spoiling. .... " 

Google and Marketing Measurement

Always have been impressed by Google's aim at better measurement, it is foundational, and not  enough attention is paid to it.  Here some of their latest:

Measurement matters: Laying a foundation for better measurement, today and tomorrow  By Babak Pahlavan Mar 2018 Data & Measurement

When we talk to marketers about their challenges and needs in digital, measurement always finds its way to the center of the conversation. We've heard from advertisers large and small that measurement on digital can be difficult and often complex. But it’s also critical to address, because effective measurement is foundational to growth.

That might sound a bit lofty, but it’s true. Better measurement helps businesses uncover the best ways to invest their limited marketing resources. Which leads to better marketing, which leads to new customers and continued growth.  

But how do you define better measurement? We’ve invested a lot of time listening to our advertisers and industry partners, and we’ve consistently heard that, to be effective, measurement solutions must be:

Trustworthy: They must be transparent and easily verified by advertisers, publishers, and third parties, including technology providers and industry standards groups.

Intelligent: They must uncover the insights that really matter to a business—which often means using the latest advancements in areas like machine learning and going way beyond simple reporting.

Actionable: They must be easy to act on, so advertisers can quickly fine-tune or change their strategy, turning metrics and insights into real business impact. .... " 

Thursday, March 15, 2018

Human in the Loop Machine Learning

Attended a very good webinar today in the DSC series.  Strongly recommend joining DSC and taking advantage of their free resources.

This Webinar answers the question you will have as a data scientist.  Where will I get the data to train my models, when its mostly held by people?

Robert Munro, CTO of CrowdFlower answers in this recorded Webinar: 

"    ... Curious about what human-in-the-loop machine learning actually looks like? Join CrowdFlower and learn how to effectively incorporate Active Learning, Transfer Learning, and Annotation Quality in your ML projects to achieve better results. 

Join us in this latest Data Science Central webinar, where we will cover the following topics:

When to use the human-in-the-loop as an effective strategy for machine learning projects

How to set up an effective interface to get the most out of human intelligence

How to ensure high-quality, accurate training data sets

How to use ML models from different domains to improve your own labeling

​This webinar will include an end-to-end look at setting up and running a job that generates high-quality training data, and shows how to incorporate that training data into human-in-the-loop machine learning systems that you can run in your own environment.

Speaker: Robert Munro, Chief Technology Officer -- CrowdFlower
Hosted by: Bill Vorhies, Editorial Director -- Data Science Central .... " 

Decision Support for Health Data

Good overview talk this morning on the complexity of gathering and analyzing health data for decision support.  With some emphasis on dementia data.   'AI' is used here as a description of the analytcs used, as well as the direct decision process.   Like that.  Plans are to have a followup talk on this topic

Speaker:  Mark van Gils:  “AI for Decision Support in Health” VTT Finland

Slides are here.   Full voice /video recording will be placed here later

Augmented Reality Sensing Form and Depth

Will this drive us to better augmented reality shopping?   At IKEA a few days ago I used some of their on floor furniture placement and design Apps, nicely done, but could have used better space understanding capabilities.  What really engages for product usage and selection  in-place?

Depth-Sensing, Algorithms And Retail Shopping Allowing AiFi To Push The Boundaries Of Interactivity

AiFi is combining artificial intelligence with mixed and augmented reality.   By Nina Salomons  in VRFocus.

Founded by former Google and Apple engineers, AiFi is combining artificial intelligence (A.I.) with ARKit on Apple products such as iPhones and iPads. Speaking to VRFocus, co-founder and CEO Steve Gu explained how AiFi has enabled consumer products to understand detailed 3D shapes and activities, including individuals and their surroundings.   .... " 

Optimizing the Usefulness of Chatbots

Still awaiting reasonably adept and useful chatbots that can do real conversation.  So below starts some key thoughts.  Tracking what the base of their knowledge looks like, and how they are being effectively used.   And how they need to be maintained.  Ultimately relevant common sense and common context will also be key to understand.

Using machine learning to monitor and optimize chatbots

The O’Reilly Data Show Podcast: Ofer Ronen on the current state of chatbots.   By Ben Lorica

In this episode of the Data Show, I spoke with Ofer Ronen, GM of Chatbase, a startup housed within Google’s Area 120. With tools for building chatbots becoming accessible, conversational interfaces are becoming more prevalent. As Ronen highlights in our conversation, chatbots are already enabling companies to automate many routine tasks (mainly in customer interaction). We are still in the early days of chatbots, but if current trends persist, we’ll see bots deployed more widely and take on more complex tasks and interactions. Gartner recently predicted that by 2021, companies will spend more on bots and chatbots than mobile app development.

Like any other software application, as bots get deployed in real-world applications, companies will need tools to monitor their performance. For a single, simple chatbot, one can imagine developers manually monitoring log files for errors and problems. Things get harder as you scale to more bots and as the bots get increasingly more complex. As in the case of other machine learning applications, when companies start deploying many more chatbots, automated tools for monitoring and diagnostics become essential.  .... " 

Wednesday, March 14, 2018

Talk: AI for Decision Support in Health - how to make it work

Upcoming CSIG Meeting:

Date and Time :  Mar 15, 2018 - 10:30am US Eastern
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
Website: http://cognitive-science.info/community/weekly-update/  (Slides, Recording)

{ Also presented at #OpenTechAI workshop in Helsinki https://developer.ibm.com/opentech/2018/01/29/helsinki-march-2018-opentech-ai-workshop/ }

Talk Title: AI for Decision Support in Health - how to make it work
Presenters: Mark van Gils (VTT)

Mark van Gils is an experienced research & development professional, specializing in data-analysis solutions for health and wellbeing applications. A successful track-record in setting-up, carrying out and managing data-analysis projects with healthcare professionals and SMEs and global companies operating in the health and wellness area.

• Impact through development of data analytics solutions that are meaningful and used in practice, impact through scientific co-operations and publications; (co-)author of over 120 articles in the field.
• Over 20 years experience in machine learning, statistics, signal processing, artificial intelligence methods.
• Leadership and management of multi-location team (>15 R&D professionals), co-ordination of large international multi-disciplinary R&D projects
• Communication of data analytics results and providing insights for different stakeholders
• Taking care of customer relationships
• Ph.D. in artificial intelligence/biomedical engineering, M.Sc. in applied physics
• Lecturing courses and guiding students and researchers


Healthcare is one of the most conservative fields in the uptake of new technologies. Reasons for this range from regulatory considerations to (informal and formal) processes that are difficult to change, but also technical issues, such as problems with the data and the difficulty of proving performance play a strong role. In this tutorial we will discuss issues we may run into when considering AI approaches for health applications. Subjects include (but are not limited to): how to get the input data right (poor quality data, missing data, harmonization), (lack of) Gold Standards and objective measures, black-box approaches vs. explainable models, data visualization, usability, classification performance vs cost-effectiveness vs practical meaningfulness. Examples of the issues and practical hints will be given based on real-life example cases of implemented systems. .... " 

Optimization vs AI to make things Better

I was reminded that my early experiences with government and enterprise systems dealt with the optimization of systems.  That is, the mathematical means of linking a specific mathematical statement of a problem, with value goals and constraints, to a specific best possible solution.   We used the predecessors of ILOG, and CPlex directly for these problems.    We saved millions using these methods.  Of course optimization does not have the current hype.

Now how is AI, as it currently defined,  dissimilar from Optimization?    Usually because the Optimization approach is more specifically and numerically defined.     If AI uses human-like intelligence, it is usually not precisely mathematical.   And unfortunately not as closely tied to specific business process.   Not saying that AI cannot use optimization methods, it just usually does not.   So there should be a strong consideration towards using more precise and direct and process oriented methods.

Was pointed to this company that works the space, have never worked with them:

Optimization Direct Inc., co-founded by Dr. Robert Ashford, a pioneer in the field of optimization, and Dr. Alkis Vazacopoulos, a leader in the industry, markets IBM® ILOG® CPLEX Optimization Studio®, the world's leading software product for modeling and optimization.

CPLEX Optimization Studio* solves large-scale optimization problems and enables better business decisions and resulting financial benefits in areas such as supply chain management, operations, healthcare, retail, transportation, logistics and asset management. It has been applied in sectors as diverse as manufacturing, processing, distribution, retailing, transport, finance and investment.

CPLEX Optimization Studio is an analytical decision support toolkit for rapid development and deployment of optimization models using mathematical and constraint programming. It combines an integrated development environment (IDE) with the powerful Optimization Programming Language (OPL) and high-performance ILOG CPLEX optimizer solvers. CPLEX Optimization Studio enables clients to:

Optimize business decisions with high-performance optimization engines.

Develop and deploy optimization models quickly by using flexible interfaces and prebuilt deployment scenarios.

Create real-world applications that can significantly improve business outcomes. ...... "

The Semantics of Image Deep Learning

Google once again shows its impressive advanced AI/Deep Learning capabilities.     Which made me recall that it is often the 'semantic', or meaning in context aspects that are most important for an AI or analytic method to be useful.   And that assigning tags also implies we will need to maintain the tags as context changes.  Below is technical, look at the link for some image examples that make this clearer.

Semantic Image Segmentation with DeepLab in Tensorflow

Posted by Liang-Chieh Chen and Yukun Zhu, Software Engineers, Google Research

Semantic image segmentation, the task of assigning a semantic label, such as “road”, “sky”, “person”, “dog”, to every pixel in an image enables numerous new applications, such as the synthetic shallow depth-of-field effect shipped in the portrait mode of the Pixel 2 and Pixel 2 XL smartphones and mobile real-time video segmentation. Assigning these semantic labels requires pinpointing the outline of objects, and thus imposes much stricter localization accuracy requirements than other visual entity recognition tasks such as image-level classification or bounding box-level detection. ... "

Assistants and Common Sense

Its actually not too often that assistants speak gibberish, they more often just admit to not knowing what was asked, when a human would understand readily.  That's usually better to diminish risk.  See my previous posts on Common Sense reasoning, which we worked on in the enterprise.  The idea of a challenge is good, it will at least scope the problem in contextual and current terms.   See the examples at the link below.  Also the notes on what 'fundamental limitations' are. 

AI assistants don’t have the common sense to avoid talking gibberish
A new test could prove that when it comes to language, today’s best AI systems are fundamentally limited.   by Will Knight in Technology Review

Siri and Alexa are clearly far from perfect, but there is hope that steady progress in machine learning will turn them into articulate helpers before long. A new test, however, may help show that a fundamentally different approach is required for AI systems to actually master language.

Developed by researchers at the Allen Institute for AI (Ai2), a nonprofit based in Seattle, the Arc Reasoning Challenge (ARC) will pose elementary-school-level multiple-choice science questions. Each question will require some understanding of how the world works.  .... " 

Wal-Mart Brings You Fresher Groceries through Eden

In Wal-Mart's blog:

Eden: The Tech That’s Bringing Fresher Groceries to You   By Parvez Musani   Vice President – Supply Chain Technology, Walmart

What’s for dinner tonight?

No matter the answer, there are some givens: It has to taste good, be good for you, and be affordable. But when you’re shopping with limited time, how can you be sure you’re buying the freshest apples, milk that will last, or perfectly ripe bananas?

We think our new intelligent food system called Eden can help. Developed in just six months by our own associates, it is improving the quality and flow of fresh groceries from farm to shelf.

Eden is the result of a friendly competition, or hackathon, among the engineers on our fresh merchandising teams. Our goal was to figure out the best way to keep track of food freshness all the way from the farms to our stores. The winning team determined that building a digital library of food standards was the answer. So they gathered the many chapters of food product specifications set by the USDA, layered on Walmart’s own rigorous product standards, and combined all of this information with more than a million photos to create a freshness algorithm that prioritizes the flow of perishable goods worldwide.  ... " 

Robert Hetu of Gartner also discusses this here:
Walmart’s Freshness Algorithm; A Great Example of Algorithmic Retailing   by Robert Hetu   

Neural Nets Remembering

Well yes, all neural nets 'remember'.  We examined that feature in their very early days.    But they don't often remember in the way we would like memory to work.    Very quickly we needed to rework their memory with new infrastructure, based on new context.  Often with new metadata.  So they remember in strict context.   Useful, but you have to be careful about the term.

Following piece is on the topic and interesting

The Neural Network That Remembers
With short-term memory, recurrent neural networks gain some amazing abilities  By Zachary C. Lipton and Charles Elkan .... 

Future of AI Assistants

Useful perspective from a number of players.

SXSW 2018: The Future of AI Assistants
Alexa, Google Home, Siri, and Cortana will learn to adjust to your changing life   By Stephen Cass

In the years to come, what will be the biggest improvement in AI-powered digital assistants? It’s likely to be the ability to accommodate a fundamental aspect of being human: The fact that we all have different personas, we show different facets of ourselves depending on where we are and who we are with, and our personas change over time. And different personas want different things from their AI assistants. Assistants that can understand your personal circumstances are less likely to remind you to pick up your rash prescription as you drive by the pharmacy if there are other people in the car, bug you about work email at home, or keep suggesting fun nightclubs if you’ve just had a baby.

That was the message from Sunday’s panel on “Designing the Next Wave of Natural Language and AI” at the SXSW festival in Austin, Texas. The panel included Ben Brown from Google; Ed Doran from Microsoft; Karen Giefer from Frog; and Andrew Hill from Mercedes-Benz. .... "