• Home

Whozat? A New Approach to People Internet Searches

An interview with co-founder Alex Baecker
bridges vol. 17, April 2008 / News from the Network


by Christoph Derndorfer

 

Recent reports indicate that up to 30 percent of Internet searches on popular search engines such as Google or Yahoo! are people related. Users want to reconnect with old friends and classmates or find out more about a potential date or babysitter, while companies are interested in any information to be found about job applicants. People are increasingly using social networks such as Facebook, LinkedIn, MySpace, and Xing to stay connected to their family, friends, and colleagues. Micropublishing using blogs and sharing photos or videos on platforms such as Flickr, Pbase, Revver, and YouTube is also becoming more popular – and not only among young people. Altogether, these trends lead to more available information about people than ever before. However it isn’t always easy to find relevant materials. Some results may be outdated, others might be useless. When searching for people with relatively common names it can be close to impossible to find the specific person you’re looking for. Therefore there is a need for specialized search engines that address this specific scenario.

 

Ph.D. Alex Baecker
Alex Baecker

bridges spoke with Alex Baecker, the founder and CEO of abInventio, which defines itself as “the invention factory.” Along with other products, abInventio has recently released Whozat? a new contender in the people search engine market.

 

{access view=guest}Access to the full article is free, but requires you to register. Registration is simple and quick – all we need is your name and a valid e-mail address. We appreciate your interest in bridges.{/access} {access view=!guest}

Alex Baecker is a native of Argentina who came to the United States to pursue his studies at MIT and the California Institute of Technology. Baecker holds a degree in biology and economics from MIT, and a M.S. and Ph.D. in computation and neural systems and biology from Caltech. He is also the founder of several start-up companies focused on search and related technologies. He was part of the ASciNA group in Los Angeles when it started and has attended meetings of the Austrian research group there.

 

bridges: What are some of the differences that you have noticed between Europe and the United States with regard to doing research and building technology start-ups?

 

Alex Baecker: The way I would put it is that when you’re in Italy, for example, be it Rome, Florence, or some other city in a European country, you notice that everybody breathes admiration for artists. People admire art and they are raised with that in mind. I think in the United States, and in California in particular, that same thing happens with entrepreneurship. The people that magazines write about and people admire are entrepreneurs who build companies. Thomas Edison and Henry Ford, for example, are classic American role models. So that permeates society because people hear and read about it all the time.

 

This mindset certainly helps, because when you’re trying to start a company, which is always an uphill battle, you need people to believe that this can succeed. The existence of role models helps people accept equity for payment early in the process, which means that employees need to be willing to work believing in what will happen in the future. While you can find investors here in California who are visionary and will invest early on, you typically won’t get investment from anyone other than friends and family unless you’re very tied into the community or have a pretty good proof that the thing you’re working on will make money. It takes a significant amount of both time and work to get this proof, and you need people to work until that point. This is certainly something that the United States, and California in particular, is very good at as people have seen it pay off well before so they’re really willing to work for a future benefit.

 

Whozat logo

bridges: One of your latest projects, which has won a lot of acclaim, is a people search engine called “Whozat?”. Within the limitations of an elevator pitch, how would you describe what Whozat? does?
  

Alex Baecker: Whozat? is a people search engine, which is a search engine that is specifically designed to find information about people. For example you can use it to reconnect with old friends and find them wherever they might be, whatever social network they are in, and if they have a blog or Web page. Or you can look for more information about a job candidate or a potential date. At the end of the day, interacting with people is one of the most important activities in life, so finding out more about them is very important to everyone. This is exactly what Whozat? is designed to do.

 

In more technical terms, Whozat? is a semantic engine, which means that it takes the meanings of words into account. It’s also an interactive engine, which means that you don’t just query and get the results. You can also tell it “this is what I’m looking for,” “this is not what I’m looking for,” or “give me more of this,” and it will change its results in real time to adapt to what you really want to find.

 

bridges: People search engines seem to be a very active field at the moment. That means that there is lots of competition, with many players in the market who are relatively young (e.g., 123people.com, PeekYou, Spock, Wink) and others who are more established (e.g., ZoomInfo, Yahoo! People Search). How will Whozat? try to distinguish itself from these competitors?

 

Alex Baecker: The basic answer to your question is quality. We have measured the relevance of the search engine results for a third-party-compiled list of young people using Google, Yahoo, Ask, Spock, Wink, Pipl, and Whozat?, and Whozat? was clearly the winner. That was even before we added interactive features that allow users to fine-tune the results.

 

Also, if you look at sites like Spock that call themselves people search engines, I would claim they’re really more of a social network, seeing that they’re mostly dependent on people for actually adding information, similar to how Wikipedia works. This approach works for famous people, as someone will add the information, but for most regular people there’s nothing there. In fact I did a search for Ashley Alexandra Dupre at the height of her 15 minutes of fame and didn’t find a single result on Spock. Whozat?, on the other hand, updates its results in real time, and as her information was changing very quickly I actually saw comprehensive results updated as the changes occurred. Today, Ryan Schallenberger makes the news for planning an attack at his school, and again Spock returns no results for him, while Whozat? returns extensive ones, including our summary right up top that he is a high school kid who got arrested, his MySpace page, a Wikipedia entry, and much more. 

 

Whozat? web site
WhoZat? Web site (click to enlarge)

bridges: In terms of the underlying search technology, there seem to be three main approaches at the moment: The traditional approach is a simple data aggregator that collects and combines information from a variety of sources such as Facebook, LinkedIn, Flickr, and YouTube. The second approach attempts to enhance and improve their aggregated dataset via a social component that allows users to edit, tag, vote, and comment on these result sets. Whozat? on the other hand is using an underlying semantic engine in an attempt to gain a better understanding of the actual keywords that are related to search results. Would you be able to shed some more light on Whozat?’s approach?

 

Alex Baecker: One of the interesting things behind it is that it’s real-time semantics. This means that there’s no database, dictionary, or ontology that knows that “car” is similar to “automobile” but maybe doesn’t know anything about “Britney Spears” which is really a more common search query than “automobile.” Our semantics come from processing pages on the Web so they actually change in real time as the meanings of terms change. So it’s dynamic but it’s also comprehensive in that it doesn’t cover a limited set of common words.


To give you an example: I was recently looking for a shop to purchase compost. I had been referred to this place that sells compost in bulk, but a search on Google didn’t turn up any relevant results. Whozat? on the other hand did find it because the engine realized that “compost” is semantically related to “fertilizer” which was the term used by the business on its Web site.

 

bridges: Via abInventio, “the invention factory” as you call it – you’re involved in a variety of innovative start-up companies and projects. What drew you to work on people search engines?

 

Alex Baecker: Try searching for John Smith in Google – you will get nowhere. This is a particularly good example to illustrate the problem. Generally, I was dismayed by how little progress is being made today in search. It’s basically the fact that major search engines, with the billions of dollars they make, are still just searching for words and texts and have no idea about the meaning of words. People search was an area where abInventio could create improvements and apply them to a problem that was particularly acute. There are many challenges facing today’s search engines, but for a start-up it’s usually better to initially focus on a smaller subset of problems. 

 

WhoZat? search results
WhoZat? search results (click to enlarge)

bridges: What are the biggest challenges that you’ve had to face while developing Whozat

 

Alex Baecker: When you’re starting something new with existing competition that has millions, as in the case of Spock, and billions, as in the case of Google, people are initially skeptical that a small start-up working out of a garage can actually do something that these companies haven’t done already. However the beauty is that you can actually prove that it works, which is very encouraging.


bridges: When did you start with the development of Whozat?

 

Alex Baecker: We started in 2006, but we were able to build on experience in developing search-related semantic technologies. The current version of Whozat? is really the result of lots of testing and input from people such as my co-founder Marzia Polito and other advisors. It is certainly a very iterative process. What is very encouraging for us is the positive user feedback, even at this private beta stage. More than half of the people who try Whozat? come back to the site within a week, so they are liking it. People are also extensively using our novel features such as the interactive fine-tuning to improve search results.

 

bridges: Whozat? is currently in private beta. When can people here in the United States and elsewhere expect the public launch of the search engine?

 

Alex Baecker: That hasn’t been determined yet. We are letting a number of people who sign up on our Web site into the private beta. The bridges readers should make sure to put “ostina” in the referral field and we might let them in sooner.

 
 

***
 

The above article is based on a phone interview with Alex Baecker, conducted by Christoph Derndorfer on April 17, 2008.

 

{/access}

 Print  Email