August 18, 2006 4:00 AM PDT

Spying an intelligent search engine

(continued from previous page)

How these services present AI search results is also very important. Google revolutionized search for the masses partly because of its simple interface, with no ads, no clutter, just a search box. For Powerset and Medstory to capture new searchers, they must help raise expectations with what turns up in the results. Technologies like speech recognition will fuel advances.

An upstart called Riya is part of the trend, too, because its technology is about recognizing more than just nouns, or keywords. Its photo software attempts to string together a sentence with a verb, such as "find a girl who looks like this girl for me on"

A match for the ratty couch
Imagine uploading a picture to the Web of your favorite ratty couch, and then asking a search engine to find another one like it. The tool wouldn't just produce a similar couch but it might even point to a store where you could buy it.

Right now, most image search engines rely on keywords, or descriptive text that is linked to a photo in order to retrieve a list of results that match a Web surfer's keyword query. That method can be unreliable, however, if photos or images lack sufficient descriptions.

Munjal Shah Munjal Shah

Riya, in contrast, looks inside the photo to extract information about its qualities using AI. The San Mateo, Calif.-based company, which launched a beta site in March, uses algorithms to calculate the photo's shape densities, patterns and textures--among many characteristics--and extract them into a mathematical representation of the photo, or what Riya's CEO Munjah Shah calls a visual signature. Each photo is represented by 6,000 numbers, Shah said, and the company uses AI to match one visual signature to another.

"All the AI is the technology to extract that information and compare two visual digital signatures," he said, without divulging the secret sauce to the technology. "It's never exact, but it's more like how do you create a fuzzy effect. That's where we've built up a lot of technology to be able to do better than most people."

The field of AI called computer vision, which encompasses facial detection and recognition, is coming of age for several reasons. One is that computing and CPUs are much cheaper and powerful now, Shah said. The dual-core computers that have come out with 8 processors have one-eighth the running cost yet are eight times more powerful, for example.

On economics and accuracy
For software to merely detect a face in a photo--not to recognize who it is--took enormous expense and time just a few years ago. Riya takes between one second to detect a face within a small photo, about a quarter of a megabyte, and 20 seconds for a photo file as large as 7 megabytes.

"To do this even a few years ago, with a billion photos on the Web just looking for a face, you would have to spend a fortune on computing, an obscene amount of money. Detection takes a lot of computing power and it's just gotten cheap enough," he said.

"The economics of it work but you still have to worry about accuracy," he added.

Recognizing an object or person in a photo is a different problem, but it's faster than detection once the computers calculate a visual signature, Shah said. Riya, for example, can compare one face to 250,000 photos in one second to find relevant matches.

Riya will be able to do many tricks, apart from couch-shopping. A Web surfer, Shah said, will be able to point to a photo of a girl on and ask the search engine to find other girls that look like her on a dating site like Similarly, someone could upload a favorite dress pattern and ask the engine to find similar patterns. Riya already offers a service called Myphotos, which lets people upload photos, train the program (e.g. "this is a picture of me and my mom") and ask it to find other photos like it in a set.

Riya, founded in August 2004, has 50 employees, including 14 specialists in computer vision. The company started out strictly in facial recognition, but realized it had a larger opportunity in object recognition. It recently raised $15 million in venture capital, bringing its total funding to $19 million. It plans to make money by selling advertising and collecting fees from merchants that benefit from Riya's visual search engine. For example, it will collect a bounty if someone ends up buying that new couch on eBay.

Previous page
Page 1 | 2 | 3

See more CNET content tagged:
Artificial Intelligence, search engine, children, word, Google Inc.


Join the conversation!
Add your comment
Artifical Intelligence is a meaningless buzz word. A rule based system would be accurate but no where near the emotional context of artifical intelligence. Inference Engine would be just as buzzy but much more accurate. I am sure others could come up with other terms. Maybe you should hold a contest to see who comes up with the best descriptive term. I submit Inference Engine.I hear AI and I instantly think Vapor.
Posted by Drewky (2 comments )
Reply Link Flag
Face recognition has great potential
Forget finding a girl that looks like my last girlfriend and will remind me of that pain. Law enforcement searches will benefit from facial recognition. From matching a missing kid to a kiddie porn image to finding a fellon that changed his/ her identity. For law enforcement it will be an new and invaluable tool, trolling the mug shots and internet in search of possible matches to all kinds of case files.
Posted by WesFlash (19 comments )
Reply Link Flag
AI is just code word for vapor/hoax ware.
I agree with other poster, AI is just code word for vapor/hoax ware.

If I had gotten a dollar for each time I heard of AI and then nothing of sort was ever delivered I would be a rich man.
There is no such a thing as AI (Artificial Intelligence) software and there wont be any, because we have no idea of how to emulate complex intelligent thinking in software, there is only good software engineering.

So it is not AI that is going to make a better search engine to Google & Yahoo, it is innovative new search engine ideas which are implemented based on good software engineering work that it is going to do it. And I will tell you about a search engine that is better than Google or Yahoo,
it is called Anoox, and it is better based on these points:
1- It is powered by the Knowledge of the people
2- It is operated in an Open fashion, so NO one company owns/controls it but 100's of different company's from around the world will
It is here in case:
Ah, another reason it is better, it is also not-for-profit.
Posted by Sea of Cortez (67 comments )
Reply Link Flag
Goolge has pushed the limit of html search
google has pushed the limit of html search. as long as pages and sites continued to be created in a way and format that does not allow for descprtive classification any real AI will be impssoble
Posted by darmik (3 comments )
Reply Link Flag
True Natural Relevance not AI
I think AI as a search descriptor is a little oxymoronic as well. The comments here demonstrate how each person has their own contextual relationship to the keyword "AI".

I do however agree that search as a science is in its infancy and that current cpu/storage efficiencies now make it possible to deliver an events-based (usage) search architecture connected to individual users, instead of the current link-topology-connected-to-no-one system weve all learned to love and hate.

Were also not ready to throw the keywordese out the door. We believe there is a lot of natural intelligence in keyword associations (i.e, John Battelles data base of intentions) that can provide an order-of-magnitude better relevance if they can be properly distilled  whether its applied to search, feeds, or media. We think the key is making it implicit (no explicit tagging or rating) and to make sure you have 100% participation  every user is both an information consumer and an information provider for every other user in the system.
Posted by Rob at Collarity (1 comment )
Reply Link Flag
Its time to roll on the next generation Search Engine
The popularity of the internet is that it offers a variety of content not available in any other medium. A search engine assists users to locate the information.

The Internet is supposed to be for fun not serious business.

And that is why most people use research tools and offline content to achieve their results. Internet search is the last place a true researcher would go to find top rated content. Only after all other options are closed.

What if the internet offered this as the very first option and people could be sure that all the information that the internet and internet search throws up is Top rated quality stuff and it is available free/subscribed/paid. I am myself a researcher and would love to have such a tool in my hands instead of having to spend a huge amount on buying proprietary research content.

Once I finished using the content, I need not pay for it. There are many websites such as which provide this sort of quality info.

Present Internet Search is neither intelligent nor smart. Cluster search engines such as Vivisimo and help to some extent but stull trash out the same stuff.

In this respect, I admire the efforts of NetAlter which is bringing a radical new search engine that would offer quality content and meta information. According to NetAlter, their search engine would offer a variety of pre-search and post search tools that enable sorting, comparing and analyzing of search results and also offer a single click ecommerce connection.

Check out the NetAlter search whitepaper and presentation.

<a class="jive-link-external" href="" target="_newWindow"></a>

<a class="jive-link-external" href="" target="_newWindow"></a>
Posted by guyfrom2006 (33 comments )
Reply Link Flag
new search required
3D search idea by Micosoft is the one i find quite useful and it invloves user interaction as well... currently many search engine do give anonymous links to increase the number of pages but only some r useful.
Himanshu joshi
<a class="jive-link-external" href="" target="_newWindow"></a>
Posted by Himanshu_Joshi (11 comments )
Reply Link Flag
AI is already being added to search engines
People need to start following the AI bouncing ball at Google and other search innovators. The integration of the CYC taxonomy, for example, appears to be well under way with available frameworks that allow natural language and human logic to prevail over "keywordism" in the very near future.

<a class="jive-link-external" href="" target="_newWindow"></a>
Posted by readyforthefuture (1 comment )
Reply Link Flag

Join the conversation

Add your comment

The posting of advertisements, profanity, or personal attacks is prohibited. Click here to review our Terms of Use.

What's Hot



RSS Feeds

Add headlines from CNET News to your homepage or feedreader.