Decisions and Technology

Ultimately its about the how technological results translate into decisions.  So that makes this interesting.   Does tech enhance decisions, and how can you make it do that?  The influence of user interface?    Organizations, like a company or even an Army are ways that have been established with ways to influence decisions internally and externally.

Does technology really enhance our decision-making ability?     By ARL Public Affairs

ADELPHI, Md. -- An Army scientist recently won a best paper award at the Association for Computing Machinery's 26th Conference on User Modeling, Adaptation and Personalization for discovering that most people cannot distinguish between liking a user interface and making good choices.

Dr. James Schaffer, U.S. Army Research Laboratory scientist stationed at ARL West, and his collaborators at the University of California, Santa Barbara, Drs. John O'Donovan and Tobias Höllerer, received the best paper award at the conference held in July at Nanyang Technological University in Singapore.

So, does technology really enhance our decision-making ability?

The paper, "Separating User Experience from Choice Satisfaction," addresses this question and furthers the theory that underpins the evaluation of recommender systems, which are designed to help users make good choices.

Simply put, recommender systems are artificially intelligent algorithms that use big data to suggest additional products to consumers based off of things such as past purchases, demographic information or search history, for example. Think of the "people you may know" feature that exists on many of today's social media platforms.

In recommender systems, it has been assumed that users form very complex mental models of user interfaces.

This is reflected in current user experience measurements, which elicit subjective responses on a wide range of system features.  .... "

Value Delivery Modeling Language (VDML)

Brought to my attention again.   Should this be the place to start for any project that claims to generate value?  Even doing a simplified model can act as a prototype from which to build from.  To test prototypes with people and other systems.  We used other BPM methods, but these were not as focused as was necessary.  Has anyone done this?  Point me to some examples.  Reconsidering.

Has been used with Watson and other AI/Expert systems for Smart Service Systems

http://www.omg.org/spec/VDML/       Also search in GitHub  …

Value Delivery Modelling Language
Value Delivery Modeling Language (VDML) is a standard modeling language for analysis and design of the operation of an enterprise with particular focus on the creation and exchange of value.

In 2009, the Object Management Group (OMG) launched a Request for Proposal (RfP) to develop a standard for value modeling. The goal of this RfP is to integrate the different existing value models and give a complete overview of the business logic of an organisation. To reach this goal, the standard has to satisfy nine requirements .... "

Purpose, use, examples of VDML  VDMBee

Use in Cognitive Computing and Watson Discover.

User Friendly and Visual Simulation Language

In the MIT News,  Faster and friendlier ... and thus more easily testable and maintainable, is always of interest.  Note the direct link to the visualization of the model.   Technical here, but I look forward to an applicable system.

MIT News (08/10/16) Larry Hardesty

Researchers from the Massachusetts Institute of Technology's (MIT) Computer Science and Artificial Intelligence Laboratory, Adobe, the University of California, Berkeley, the University of Toronto, Texas A&M University, and the University of Texas have developed a programming language called Simit. Simit can accelerate computer simulations by a factor of 200 or reduce their required code by 90 percent, according to a paper the researchers presented last month at the ACM SIGGRAPH 2016 conference in Anaheim, CA. 

"The story of this paper is that the trade-off between concise code and good performance is false," says MIT graduate student Fredrik Kjolstad. "It's not necessary, at least for the problems that this applies to. But it applies to a large class of problems." Simit requires the programmer to describe the translation between the graphical description of a system and the matrix description, but following that the coder can use the language of linear algebra to program the simulation. 

During the simulation, Simit does not have to translate graphs into matrices and vice versa, instead converting instructions issued in the language of linear algebra into the language of graphs, keeping the runtime efficiency of hand-coded simulations. Kjolstad notes Simit-written programs can run on either conventional microprocessors or on graphics processing units with no underlying code revisions. 

Encyclopedia of Interaction

An update on this work, which I first examined in 2011.  Still apparently a free resource.   The Encyclopedia of Human-Computer Interaction, 2nd Ed. ...  from the  Interaction Design Foundation. which now also sells courses in the area.  

A Designers Guide to AI

AI from a design perspective.   Most often these days see it from a engineers or data science view.   We should also consider it from the point of view of what the decision maker needs.   Good lengthy piece.

" ....  designers know the basics: The principles of user-centric design lay the groundwork for building a great AI system. The approach is rooted in anthropology, and it has taught designers to do a lot of things well. The best designers know how to create things that people want, and understand how, to use. But until recently, that’s where the design process stopped.  ... " 

by Patrick van Hoof   ... With number of design oriented links.

Using Behavioral Data

In Linkedin by  Robin Cronan

Using Behavioural Data
Marketing Datafication No.2 looks at behavioural data and how it can be used to improve targeting, message personalisation and customer experience ... " 

Facebook Opens Paris AI Center

In TechCrunch:  Facebook: More than just identifying faces, using information in social to create a number of alternative models and link them to engagement and value.  AI is back.

More on SocioPragmatic Text Models

Yesterday posted about the talk on SocioPragmatic text Modeling.   More has been published in the Cognitive Systems Institute Linkedin Group,    I quote some of my recent input below   This is a new kind of text classification modeling based on consumer cognition .

" .. Also mentioned in this talk, the potential use of this approach for market research. Suppose a subgroup is the collection of likely buyers of your product? Then classification based on this approach would predict sales. This is another example of 'non conscious' classification, like methods called 'neuromarketing', which often use electroencephalographic techniques, being used by some companies today to augment classic market techniques. This would be far cheaper to implement on more people, even in mobile environments. Using this with multiple languages could provide global reach, also a plus. Consider its use with Twitter streams.  There is real potential here. ... 

This Sociopragmatic method also relates to the 'User Modeling' Watson capability, now available free to test on Bluemix. That method claims a psychological, rather than social-linguistic classification basis. A comparison of the two methods value would be very interesting. Say with the same Twitter stream? Will bring this up with a retailer or manufacturer.  ... "

Hope there will be more input there, feel free to connect and join in.

Surprise and Simplicity in User Experience

Discussion in Retailwire:   Good thoughts on the retail user experience.    It should be about delighting the consumer I was always told, but the balance with simplicity is also important:.   From the Lenati blog:

" ... As customer experience strategists, we are seeing two popular competing approaches to customer experience (CX)—“surprise and delight” versus “effortless simplicity.” These are both valid CX approaches depending on the customer’s context, and this has led to a fair bit of confusion around how people speak about CX design.  At Lenati, we feel it’s time to clear that up  ... "    (Good thoughtful read)

Failed Personalization

Have been involved in a number of projects that were aimed at customizing response to consumers based on a technical segmentation or personalization.  I agree that many of these attempts are poorly done. Look at some of my recent notes on the use of technical user modeling.  Discussion in Retailwire. 

" .. Consumers aren't too happy with the "personalized" offers and communications they're getting from retailers. A recent survey found only 22 percent of consumers believed the average retailer understands them as an individual. Why are retailers' personalization efforts getting such a low rating from consumers?  ... " 

Twitter Services on BlueMix Creating Collective Intelligence

Results from the collaborative work between IBM and Twitter.  Relates to other things I have posted here recently on User Modeling.    To what degree does Twitter act as 'Collective intelligence'?    will continue to follow.

  " ... Last October, IBM and Twitter announced that they were combining forces to integrate Twitter data with IBM analytics services on the cloud that will help transform how businesses understand their customers, markets and trends – and inform every business decision.

Today, we’re excited to announce the availability of IBM Insights for Twitter service on Bluemix. Developers can now take advantage of the collective intelligence of hundreds of millions of people to create unique, social media applications quickly and easily, directly in the IBM Bluemix environment. The service enriches tweets, based on deep natural language processing algorithms from IBM Social Media Analytics. Real-time processing of Twitter data streams is fully supported; configurable through a rich set of query parameters and keywords. ... " 

Juji: Behavior-based Individuality Analysis

Michelle Zhou brought to my attention their new personality analysis startup that creates people profiles from Twitter and Facebook streams.  I  gave it a try and was impressed by its ease of use.     The idea of this kind of user modeling has been around for a long time, we tested related ideas with text analyses of 800 line calls.  This example is one of the first that uses new sources of social data effectively.  Consider if you combined results of this system with streams of IOT data, you could provide an analysis of people in context.

Juji has similarities to and in part derives from work done at IBM, that I have reported on previously, but it is now a unconnected startup.  It provides both profiles and alternative use case scenarios.  Give it a try, they need a critical mass of testers, and are providing incentives for students to participate.

Sign up and test the system.

They write about their goals:

Our Vision
Juji's mission is to discover the world’s people by their individuality, and to enable individualization at scale. Our notion of individualization includes two aspects. One is self individuation, a concept proposed by the renowned Psychologist Carl Jung: one should be aware of and realize one's true self. The other is for individuals to obtain hyper-personalized experiences that suit their true nature, including their personality, physical and emotional needs, values and beliefs, and taste and style.

As the first step toward fulfilling our mission, we empower our users, individuals and businesses alike, to automatically derive deep insights about people of interest from opt-in or public data (e.g., public blogs and web pages). Such insights characterize a person's or a group's individuality from multiple aspects, encompassing their professional and personal lives. Our users can then use the derived insights to develop and improve their own brand personality, and create individualized engagements at scale (e.g., interacting with hundreds or thousands of customers in an personalized way).  ....    "

Building User Models with Twitter

Have recently had to look at how user models can be built from a Twitter stream.  In particular for retail interactions.  Took a look at some of IBM's Watson influenced services in Bluemix.  Here a non technical piece in the New Yorker on the topic. The question remains if you can create a user model that has some clear value.  Good piece.

Content Engineering and AI

Short piece by Omri Astel made me think about how ad content can be engineered based upon the audience.  Early steps have been taken in this direction.  For example, with Watson User Modeling. See this Description/Documentation.    And also some demonstrations.   We experimented with similar methods in the late 80s, managing the interactions of our brand equity.  Based on those early experiences, considerable progress has been made, but not yet to the level of what we would call intelligence.

User Centered Design

In Wired:  Fascinating piece.   " ... How the Home Telephone Sparked the User-Centered Design Revolution .... Have you ever thought about why doorknobs are positioned at around two-fifths of the door’s height, instead of right in the middle? Or why a washing machine is of its particular shape and size? These sound like stoner thoughts, but there’s actually an entire discipline and rich design history devoted to the answers. ... "  ... " 

Personality Judgements by Computer

Supporting information for a proposed project:
" ... This study compares the accuracy of personality judgment—a ubiquitous and important social-cognitive activity—between computer models and humans. Using several criteria, we show that computers’ judgments of people’s personalities based on their digital footprints are more accurate and valid than judgments made by their close others or acquaintances (friends, family, spouse, colleagues, etc.). Our findings highlight that people’s personalities can be predicted automatically and without involving human social-cognitive skills. ...  " 

Interactive Intent Modeling

A natural interaction of user and their behavior.  Note a number of recent posts in a related topic that uses personality modeling.  Here a hybrid approach.  Currently examining this with a related tech.

In the CACM:
Interactive Intent Modeling: Information Discovery Beyond Search
Combining intent modeling and visual user interfaces can help users discover novel information and dramatically improve their information-exploration performance. .. "

" ... We introduce interactive intent modeling, an approach promoting resourceful interaction between humans and IR systems to enable information discovery that goes beyond search. It addresses the vocabulary mismatch problem by giving users potential intents to explore, visualizing them as directions in the information space around the user's present position, and allowing interaction to improve estimates of the user's search intents. ... " 

Google Releases Purchase Prediction Tool

In Mediapost:   Most interesting idea, especially as it deals with a perhaps unique path to conversion. Very interesting problem from an analytical point of view.  Includes where interactions occur on the journey.  Read the rest of the article, more details and statistics.  Details of the data being used to derive the prediction is unclear.  [ Update: The link below mentions the data used]  A form of retail oriented user model?  Looking further, hope to find out more.

" ... In an attempt to better understand the path to purchase, Google released an online tool that enables marketers to predict the journey in which customers will take to make a purchase. The tool, The Customer Journey to Online Purchase, allows marketers to explore for businesses how any of the 14 industry channels influence purchase decisions.

The tool enables marketers to pick from 14 channels including Arts & Entertainment, Fitness & Beauty, or Law & Government, as well as the size of the business, small, medium or large, in a variety of countries such as the U.S., Brazil or Japan. For each market, channels such as social or paid search are posted toward the left of the graph based on the country and business size. ... " 

Building User Models with Twitter

An outline about how Bluemix Watson services can be used with Twitter data to do User Modeling.   With a fairly simple process.  This is ultimately quite a powerful idea.  Targeted advertising and loyalty engagement.   Create an automated user model based on online behavior.  It remains to be seen how accurate it can be.  More to follow here.

Watson Determining Personality and Resonance of Message

Today at the Insight meetings we saw a Watson cognitive App being built from a number of Bluemix services and prefabricated Java.  Of particular interest was a user personality modeling mechanism.   They showed it being used to determine how well a message resonated with a recipient. A near definition of advertising.

The simple model was built and embodied in an App in a few minutes.  What they showed could be done by anyone with just a little computer dev experience.  Potentially a very big idea, especially when being used online.  Genesys was an example of a company using the idea today, for call center style applications.

It was unclear what theory of personality was being used.  Someone know?  I am sure there will be some push back on relevance of the model.  Also matches well with service and advisory applications.   I have asked for a complete demo video of the personality feature and will post it here, hopefully sometime soon.

#IbmInsight