Bill Roth, Ulitzer Editor-at-Large

Bill Roth

Subscribe to Bill Roth: eMailAlertsEmail Alerts
Get Bill Roth via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Related Topics: Java Developer Magazine, Consumer Electronics, Telecom Innovation

Blog Feed Post

Star Trek, the Search for Context

As a child I was mesmerised by Star Trek (the original series of course) as I was fascinated by thinking about all the great gadgets we would have in the future. Now that that future has arrived I see communicators are an everyday reality with the advent of mobile phones, we also have the technology for laser guns and early teleportation! But, the elusive universal language translator still evades us.

When I first started getting interested in Artificial intelligence 20 years ago I was surprised to read about the Turing Test and how difficult it had proved to pass. This test was designed in 1950 as a means of defining if a computer was actually ’intelligent’ through an ability to fool a human into thinking they were having a text based conversation with another human. Sixty years on from the birth of this test it is still an elusive test to pass. Why is this?

The fact is that despite Moore’s law and continued advancement in computing power, human intelligence still exceeds the ability of computers in tasks that require the communication of ideas, understanding them correctly and reacting appropriately. There are a number of reasons for this:
a) richness of language (terminology): there are many ways to express the same concept, for example, as a lay person I use the term heart attack, whereas a Physician is more likely to use the term myocardial infarct. Computers cannot easily make these associations.
b) a common understanding of how the world works (general background knowledge): this is why so much time (and expense) has been invested in developing taxonomies and ontologies to provide computers with a model of the world to work from
c) an understanding of un-communicated implicit information (Context): so when I talk about java am I talking about coffee, horse racing, Indonesia, prehistoric man, or web programming?

These are all underlying principles behind translation which is why the universal translator has proven so elusive to date.

When it comes to computer aided searching of content whether it be enterprise search or internet search, many of these same challenges apply. When a user submits a query in a search box what do they mean by that query? Do they just want a list of documents back that contain their search term (I think not) or do they want a more Intelligent Search that also includes all those documents with conceptually similar terminology (eg heart attack, heart disease, myocardial infarct)? Do they also want to have the ability for the engine to understand that when they enter their query they are interested in the prevention of heart disease as opposed to the causes of it or even surgical techniques to treat it, in order to receive more contextually relevant results and less noise? In the same way as translators need to understand terminology and the context of what the user is talking about to give an accurate translation, search tools need to do likewise to provide the most relevant information to the user’s needs at that time.

With Sophia we have invested over 30 person years developing solutions to search problems such as these. Our motivation was to solve these problems without relying on Taxonomies, Ontologies or dictionaries. We felt strongly that it was important to crack the elusive context problem in search and have built our technology on a model of linguistics called Semiotics – the science behind how humans understand the meaning of information in context. By improving the computer’s understanding of context, we increase the richness of communication between man and machine and make it possible to return more useful search services as a result.

After all if humans can master time travel then contextual search shouldn’t be that surprising!

Read the original blog entry...

More Stories By Bill Roth

Bill Roth is a Silicon Valley veteran with over 20 years in the industry. He has played numerous product marketing, product management and engineering roles at companies like BEA, Sun, Morgan Stanley, and EBay Enterprise. He was recently named one of the World's 30 Most Influential Cloud Bloggers.