Skip to Content
Uncategorized

Searching for Humans

Various websites are trying to make it easier to find friends and colleagues online.
August 20, 2007

Jaideep Singh, cofounder of the new people-search engine Spock, says he wants to build a profile for every person in the world. To do this, he plans to combine the power of search algorithms with online social networks.

Singh says he got the idea for Spock while looking for people with specific areas of expertise among his contacts in Microsoft Outlook. Although he has two or three thousand people listed, he could only find people he was already thinking about.

Spock is designed to solve that problem by allowing users to search for tags–such as “saxophonist” or “venture capitalist”–and then view a list of people associated with those tags. Singh could have manually entered tags for each of his contacts into Microsoft Outlook, but capturing every interest of each particular individual would be time-consuming. Spock uses a combination of human and machine intelligence to automatically come up with the tags: search algorithms identify possible tags, and users can vote on their relevance or add new tags. Registered users can add private tags to another person’s profile to organize their contacts based on information that they don’t want to share. For example, a contentious associate might be privately labeled as such.

The social-network component of the website introduces an element of crowd commentary into the search process. George W. Bush is tagged “miserable failure,” with a vote of 87 to 31 in favor of the tag’s relevance as of this writing. Users aren’t allowed to vote anonymously, and the tag links to the profiles of people who voted.

Singh hopes social networks will also help with one of the main problems in people search: teaching the system to recognize that two separate entries refer to a single person–a problem called entity resolution. For example, a single person might have a MySpace page, a Linked In profile, and a write-up on a company website. Steven Whang, an entity-resolution researcher at Stanford University, says that there are several aspects to the problem: getting the system to compare two entries and decide whether they are related, merging related entries without repetition, and comparing information from a myriad of possible sources online. Finally, Whang says, there is a risk of merging two entries that should not be merged, as in the case of a name like Robin, which is used by both men and women.

Many of the people-search engines try to get around these problems by encouraging people to claim and manage their own profiles, although Whang notes that this is a labor-intensive approach. Although there are many sites where people could claim their profiles, Singh says he thinks one engine will eventually dominate, and people will make the effort to claim profiles there. Bryan Burdick, chief operating officer of the business-search site Zoominfo, says that 10,000 people a week claim their profiles on Zoom, in spite of having to provide their credit-card numbers to do so.

Singh has also introduced the Spock Challenge, a competition to design a better entity-resolution algorithm. He says that 1,400 researchers have already downloaded the data set, and they will compete for a $50,000 prize, which will be awarded in November.

But although Spock and Zoominfo both stress the importance of being able to search by job title or other keywords, Michael Tanne, CEO of people-search engine Wink, says that isn’t the most common need in people search. “That’s not how 90 percent of searches are done,” he says. “When you search for the iPhone, you want to see what’s out there about the iPhone. But when you search for a person, you have a specific result in mind.” While his site does allow users to search by tags, Tanne says that the tags are more commonly used to narrow down a search. Wink is able to search by variables such as location with more focus than the simple word recognition Google uses, Tanne says. For example, he notes, Wink would recognize that Framingham is close to Boston, and it would include both when a user enters “Boston” as a search term. (A spokesperson for Google says the search-engine giant currently has no plans to develop special features to improve people search.)

Singh says that Spock has indexed more than 100 million profiles so far–a reasonable start on the way to indexing every person in the world. But some people have raised privacy concerns. The people-search engines spider business sites and public Linked In profiles, but also social-networking sites such as MySpace. (Facebook information is kept private.) Business and personal information can appear on the same page, although Zoominfo attempts to index only the former. The sites’ spokespeople all agree that the burden falls on the user to watch out for what’s available online. “Whether you’re managing it or not, you have a digital persona,” Burdick says. Wink will remove profiles upon request, although not all people-search engines share that policy.

“The ability to see what you’ve put out there is an eye-opener to most people,” Tanne says.


Credit: Spock

Crowd commentary: People-search engine Spock uses a combination of algorithms and user input to rank results and tag people with relevant keywords. The results can be quirky, as above: a search for “saxophonist” returns Bill Clinton as the top result, above jazz giants John Coltrane and Charlie Parker.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.