Tuesday, September 11, 2012

What in the World……

What in the World……

          Is happening with Information search and acquisition?

(NOTE: I am starting a series of blog posts dealing with interesting developments in the world of information technologies.  Entitled “What in the World...” [WitWo], this series will present my views of important trends and changes affecting all of us.  While I don’t claim any profound ability to provide insights, I do research and digest many related materials in my quest to understand and accommodate the exponential developments in Information Technology.  Hopefully, I can provide both the ‘what’ and some of the ‘so what’ aspects in this series.)

About two years ago I (along with my son) wrote a chapter in a book (Knowledge Management and E-Learning), entitled “From Self-Service to Room Service: Changing the Way We Search, Sift, and Synthesize Information”.   In this chapter, I outlined, with some early examples, how the way in which we deal with information is changing.  Although not claiming to be prescient, I think many of the trends we outlined in that chapter are now starting to take hold.

Background: Up until now, obtaining, organizing and incorporating the necessary information for actions or decision-making has been solely left to the individual efforts of the user.  The user searched a finite information source (a library or database), found and extracted the needed information, and then synthesized it into the proper format (e.g., white paper, term paper, report, action paper, etc.).  Within this “self-service” model, the user has the burden of finding the proper information, filtering it into the proper format, fusing it into a credible document relevant to the need, and distributing that document to the appropriate parties.   In an era where information was relatively limited and access to that information mainly local (e.g., the local library or a database within the organization), this model worked well for many.  The user simply found whatever information he or she required and synthesized it into the necessary form.  Certainly, modern search engines (e.g., Google, Bing) have made the physical effort needed less (i.e., no longer have to visit a brick-and-motar library) and the information resources easier to access, but the task has basically remained the same----find the exact information desired, put it into the proper context, and incorporate it where needed (i.e., “self-service”).

However, with an information hyperabundance, the tasks of searching, sifting, and synthesizing information may no longer be manageable by human effort alone.  Indeed, the “self-service” model of information retrieval and digestion might necessarily have to change to a more of a “room service" model, wherein alternate approaches to dealing with information have to be explored and, where useful, adopted.

In the “room service” model, the user requests what information (contextually based) he or she needs, in what format the information is desired, when the information is needed and the way in which the information should be delivered.  Then, in an effective “room service” approach, the information arrives at the proper time, in just the right amount, in the proper format, and contextually relevant to the problem or request at hand.  In essence, the user is ordering information to be prepared in advance, not unlike “room service” in a hotel.  If the “room service” is successful, the user will be provided with the right goods, in the right format, at the proper time.  “Room service”, properly implemented, relieves the user of some of the burden of searching, sifting, and synthesizing the information.

Two years ago many of the examples I provided in the book chapter were early, cursory efforts, including applications such as Hakia, Cha-Cha, and Collecta.  These early efforts attempted to provide some of the aspects of the “room service” model outlined.  Although they are admirable efforts, they are not close to approaching a complete solution.  Now it appears that we are much closer to a full implementation of the “room service” approach.

Current Developments: 

Several key developments in the last year provide evidence that the “room service” approach to acquiring and incorporating information is rapidly emerging.  Advances illustrating the tremendous progress include:

·         Watson:  Likely the most impressive emergent approach to acquiring and synthesizing information automatically is Watson.  If you have not viewed the performance of Watson on the Jeopardy show in February 2011, you should view it to get a feel for just how impressive this breakthrough is.  Watson is able to rapidly search extremely large volumes of information, find the requested item, format it into a natural language answer and deliver that information.  Further, Watson is able to deal with nuances and subtleties better than any previous computer program developed.  Clearly, the many uses planned and being implemented for Watson (e.g., finance, medical) evidence the value of this approach.  The next step will be a Watson-like interface for the average consumer via tablet or smart cellphone.


·         Siri:  First demonstrated as a standalone project in 2009 and now part of the Apple family, Siri has a unique ability to process requests and deliver natural language complete answers to many different types of queries.  Siri is now incorporated into the Apple iPhone 4GS and constantly improving.  For many, this development represents the initial foray into having an ‘intelligent assistant’ ready to provide instant information.  As with Watson, Siri is likely to improve rapidly (e.g., adding Yelp and Wolfram Alpha to its personal assistant data repositories) and also be embedded in many more areas (e.g., iPad).


·         Wolfram Alpha:  Wolfram alpha, the amazing work of Stephen Wolfram, is another product which is altering how we obtain and incorporate information.  Unique in many ways, Wolfram Alpha also provides answers to queries in a more complete manner vice just providing links to pages which might contain the needed answer.  Check out the examples of Wolfram Alpha in action to see just how varied and complex queries can be and still be addressed by the software.   Although primarily focused on quantitative areas, Wolfram Alpha is spreading into many other arenas.  (NOTE: Wolfram Alpha is linked to and integrated into Siri, adding the abilities of the Wolfram Alpha engine to the Siri intelligent interface, thus boosting the capabilities of both.)


·         Google “Knowledge Graph”: Not to be left out of the evolving way in which we search, sift and synthesize information is Google.  The search giant is also well into developing a more robust way of providing information.  Within the past few months, Google has introduced its “knowledge graph” concept for representing information and information linkages.  This knowledge graph contains 500 million entities liked by tens of thousands of different types of relationships (including relationships between people and places, things, etc).  With the knowledge graph, Google can now provide more complete answers to queries, including more focused context.  By using the knowledge graph relationships, the Google search engine can automatically make connections of value, based on inter-relationships.  Similar to how Siri is used, the new Google approach is now embedded in Android smartphones.


·         Microsoft Satori: Perhaps the least known development in making the web better understood (so as to provide more complete answers) is the Microsoft effort, Satori.  Like Google, Microsoft has compiled its own knowledge graph containing 350 million entities.  Incorporating Satori into the Bing search engine will provide users with much more context relevant search results.  Also, like Google and Apple, the improved search engine approach will be a prominent feature on mobile devices.


·         Others: There are other efforts also on the horizon or already underway.  One of the earliest reasonably successful efforts is Ramona from the futurist Ray Kurzweil.  Chatbots, like Ramona, add the dimensionality of a persona to the interface.  By providing a knowledge navigator interface (see the original “Knowledge Navigator” vignette from 1987), chatbots are also another way in which information is being provided in a more natural manner.  Likely, clever and realistic human interfaces will be linked with the knowledge repositories mentioned in some instances.

Obstacles and Caveats:  As with any technological development, there remain some problems to be confronted.  To me, the key obstacles are not the technology itself but instead two other related ones, namely:

·         Expectations: My experience with many information technology developments is that the expectations of the user always outstrip the ability of the technology.  We expect solutions that are ‘absolute’ rather than ‘relative’.  Absolute solutions are free, complete, exactly right at all times, etc.  Instead, we see solutions (many from the field of Artificial Intelligence) that are ‘relatively’ better than what we have, but not ‘absolute’ solutions.  For some, this delays and negates their use of the technology (laggards or late majority adopters according to the “crossing the chasm” model).  For these people, the solutions will likely never be acceptable.


·         Privacy and Security concerns: For an intelligent agent to be able to provide answers and assistance “just right” for you means they must have an excellent grasp of who you are, what you like, what your interests are, etc.  Thus, the agent will have an extremely intimate knowledge of you as a person in order to be effective.  Want an agent that intelligently and automatically pays your bills?  The agent needs to know your bank account numbers and passwords.  Would you like your agent to automatically shop for and purchase birthday and anniversary gifts for loved ones?  Then, the agent needs access to your credit cards.  If that agent is compromised in some way (perhaps identity theft in the future will include the theft of your personal agent!), then likely some (much?) of your personal information will also be compromised.  For an example of just how much information will likely be known about you, view this short vignette on ordering a pizza in the future, http://www.youtube.com/watch?v=RNJl9EEcsoE.


Conclusion: So, this leaves us on the cusp of having an “intelligent assistant” at our beckoning, available 24/7, contextually aware, and able to provide us complete information on almost any query.  As this new approach continues to evolve, we will soon approach the metaphor contained in the landmark article I have long used as an example, “The Semantic Web”.  Some will argue that we are still not close to having a complete intelligent assistant.  While that may be true, the exponential progress evident in the last few years leads me to conclude that within the next few years, we will certainly arrive at a point where an “intelligent assistant” is available and essential for all of us.