Surf Canyon, a developer of internet search technology, today announces the availability of its Discovery Engine for Search at www.SurfCanyon.com. As the quantity of content on the internet expands exponentially, it is becoming increasingly difficult for searchers to find the information they need amongst all the irrelevant results.

By using Surf Canyon’s semantic real-time implicit personalization technology (available as a browser extension for IE and Firefox) while searching on Google, Yahoo! or MSN, searchers are able to more quickly and easily find relevant information that’s buried within the usually overwhelming quantity of search results. Rather than leaving users stuck with a static set of links, the product uses real-time behavior signals to calculate the “instantaneous relevancies” of the documents in the result set in order to dynamically bring forward the most pertinent while pushing back those that are irrelevant. The technology works similarly well for targeting sponsored links.

“The concept was born out of frustration,” said Mark Cramer, Surf Canyon CEO. “While many things are certainly easy to find, whenever I was looking for something mildly obscure, I felt that I could benefit from a search technology that would work with me during the process, as opposed to leaving me the entire onus of digging through the results and reformulating the queries. Our product mitigates this problem by transforming the search process from a lecture into a conversation.”

“Surf Canyon significantly improves my search experience and the deeper results it delivers are spot on virtually every time,” said Carla Thompson, senior analyst with The Guidewire Group. “I also really like that it functions within my existing search; once installed, I’m able to enjoy the advantages without having to adjust my habits.”

Here is an example of Surf Canyon’s improved Google results:

Just for reference, here is a French post written before today’s update, with a link to the whole article, and the English translation below (thanks to Mark).

Surf Canyon aiguise la curiosité et approfondit la recherche Par Alain Joannes

Entre le moteur qui maintient la collecte à la surface des océans d’informations et le métamoteur qui fait perdre du temps en déversant trop de résultats peu pertinents, Surf Canyon inaugure une approche prometteuse de la recherche. Il s’agit d’un outil en ligne qui interprète la curiosité du journaliste en temps réel et de manière évolutive.

The whole post is here.

Here is the same post in English:

By: Alain Joannes

Between the search engine that maintains a collection at the surface of the oceans of information and the meta-search engine that wastes time pouring too many irrelevant results, Surf Canyon inaugurates a promising approach for research. It involves an online tool that interprets the journalist’s curiosity in real time in a manner that evolves.

Schematically: the more a journalist expresses an interest in one of the top results, the greater the application’s ability to “understand” what the user is looking for and then propose results in the same vein that the journalist is trying to find. Inversely, if the journalist does not express an interest in one of the results, the application “understands” that that theme does not interest the user and then avoid proposing any tracks that have to do with that theme. The user can certainly express an interest in all of the top results and he can even utilize successively five different search engines for the same query.

In the example above, the top six results are A to F. If I express an interest in result #3 (C), Surf Canyon “understands” that certain words in that result attract my attention and will then go to level -1, in the immense collection of search engines with which it’s associated, and find six more results, G to L. My interest is now in certain words, or in a particularly credible source, in result #4 (J). The application assembles the semantic “symptoms” of my interest for result J. Then, using the singularities of J, digs even further, deeper, in order to “bring forward” from level -2 the results M, N and O. And so on until – in this example – the source P, extracted from level -3. A source rare and precious above all.

First advantage: I did not waste time with 13 results A, B, D, E, F, G, H, K, L, M, O, Q.

Second advantage: my research benefits from the quality, that is to say the pertinence, of an assisted piloting by a form of embryonic artificial intelligence.

Third advantage: this form of rudimentary semantic intelligence does not prohibit me, to the contrary, from staying in control of the different orientations of my curiosity. It is sufficient to abandon one layer and go to another to dig elsewhere, not far from the present seam, but in perhaps a richer zone.

I tested Surf Canyon using the same hypothesis as of last October 8th with Quintura: “What must one think about the idea that the United States could launch a military operation against Iranian nuclear installations?”

Surf Canyon “understood” that one of the 13 top results intrigued me and went looking for neighboring results:

In addition to the deepening and diversification of sources, the “intelligence” of Surf Canyon bring me a discrete aide for understanding. In the complementary results offered above, the word “possibility” is replaced by the word “chance.” This semantic sliding intrigues me. I click and discover the point of view of the Guardians of the Revolution, interesting data for my evaluation.

During this exercise of verification of a hypothesis, Surf Canyon brought me:

· Three extremely precious sources, valuable for other geopolitical research.

· A surprising chronological reconstruction of the hypothesis “The US attacks Iran,” with all the alternatives.

If one day the hypothesis is verified, I already have at my disposal an enormous amount of documentation filled with passionate illuminations. If the hypothesis is not verified, I already possess the material for the analysis of the origin, propagation and deformation of that which will have been a geopolitical phantasm.

In my opinion, Surf Canyon gives the best of its potential in preventative research, when the journalist has the time to collect the data for a case or event that he envisions will happen suddenly.

As soon as possible, I will test Surf Canyon in a situation of journalistic urgency.

<–end–>

And I encourage you to install Surf Canyon today and try it out, then come back and comment!

For a detailed discussion of how Surf Canyon solves the Search problem:

Background:
Search Engines are the primary tools for information retrieval on the internet. A search engine attempts to determine a user’s information need based on short (typically 2- to 3-word) queries and delivers a ranked list of the most pertinent documents to the user.
Early algorithmic search engines such as Excite and Alta Vista relied solely on the document content to estimate their relevancy for a given query. Google revolutionized internet search by using the link structure of the web to computationally determine the quality of documents. As the web grows and computers become more powerful, commercial search engines have added many new inputs to determine document relevancy, including analysis of similar queries by all search engine users, neighbors (geographic targeting), friends (social search engines), and the individual’s search history (personalization).
Surf Canyon, however, changes the very nature of the search results page. Rather than being a static list of documents deemed by an algorithm to be the most relevant for a query, the search results page becomes an active participant in the process of resolving the user’s information needs. Based on the user’s actions, Surf Canyon infers intent in real time, dynamically updating the search engine results page.

Problem:
It is becoming increasing difficult for users to find the information they need with only traditional search technologies. While the amount of data on the internet expands exponentially (the major search engines now have indices that contain tens of billions of items), the amount of information entered for the average query (two to three words) has remained relatively steady. Mapping three words to billions of items is a profoundly difficult, if not mathematically impossible, task.
It is for this reason that we see studies like these:
“People search for 11 minutes on average before finding what they’re looking for, and half abandon searches without getting that far.” – Microsoft
“Half of potential Web sales are lost because visitors simply can’t find what they want.” – Gartner
“According to Outsell, in a study indicating rising frustration with internet search, ‘in many instances the culprit is search failure due to irrelevant results.’” – MediaPost, June 2006, “Irrelevant Results Threaten Search”
“79% of professionals feel that their queries are not always understood and only 10% find what they are looking for on the first attempt.” – Convera
Something more is needed.

Solution:
Many new search technologies are working to solve this problem, including clustering, personalization, tagging, voting, social networking, natural language processing as well as meta-search, new visualization techniques and others. However, all of these techniques look at the search results page as the final product. Once produced and delivered, the onus is upon the user to dig, potentially find something useful and, if unable, reformulate the query.
Surf Canyon considers the search result set to be dynamic. The first ranking of results might have been the best guess given the knowledge that was available at the time the user executed the search. However, as soon as the user takes his or her first action, that ranking quickly becomes obsolete.
By observing the user’s behavior as the search is taking place, Surf Canyon’s technology builds a real-time model of inferred intent, which is then used to calculate the “instantaneous relevancies” in the result set. The instantaneous relevancies are then used to immediately move forward the most pertinent documents while pushing back those that are irrelevant.

This entry was posted on Tuesday, February 19th, 2008 at 8:00 am and is filed under Alts, News. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.