Need autonomous driving training data? ›

Using Crowdsourcing for Complex Data Problems: It Can & Should Be Done!

Using Crowdsourcing for Complex Data Problems: It Can & Should Be Done!

Crowdsourcing companies came onto the scene with the goal of helping businesses solve simple problems with massive scale. A decade ago, we couldn’t rely on a computer to, for example, identify duplicate product pages or tell the difference between fruit and faces in an image, so crowdsourcing platforms became a smart way to disperse these easy-but-not-so-easy-a-computer-can-do-it data tasks to a large group of humans—a crowd—to tackle. Of course, today, smart machines are able to handle those tasks—faster and at better scale.

But if this old model—anonymous members of a crowd completing rote tasks—is what you think of when you think of crowdsourcing, you’ve missed a major development: intelligent crowdsourcing. With intelligent crowdsourcing, you get the good of traditional crowdsourcing (human insights, speed, scale), with the added benefits of community (not crowd) and support for complex (not just simple) problems.

Before we go on, when we say “complex” data problems, we’re talking about stuff like getting high-quality labeled data to train your machine learning algorithms, enriching your data via subjective human insights to optimize your search and browse experience, and completing and verifying key business data. Stuff that, in order to be done right, requires taskers with specific skills, expertise, demographics, etc. On that note…

The Problem with Anonymity

Crowds are largely anonymous. Anonymity is a problem because it invites cheating and sub-par performance. It’s easier to rush through tasks or scam the system when one’s name or reputation isn’t on the line. Crowdsourcing customers constantly complain of fraud, bots, and low quality—all undoubtedly a result of, or exacerbated by, anonymity.

We can look to academia for best practices on how to apply algorithms, workflows, and game theory to improve results, but the problem remains: lack of clear, personal identity means lack of accountability.

The Opportunity With Community

Members of a community have identities, communicate with each other, feel connected, feel valued, and feel accountable. Accountability = quality insights. Identity = profiles. All people have unique demographics, specialities, skills, interests, and performance reputations. For a crowdsourcing solution to help solve complex problems, it must know, track, and empower the use of its community members’ individual qualifications.

Intelligent crowdsourcing accomplishes this. With the ability to target very specific taskers, companies can get the high-quality, specialized insights they need to solve tough, complex data problems.

Pretty exciting stuff. Wanna learn more? Get the details on intelligent crowdsourcing or learn how it works.

image credit: nerovivo via CC BY-SA 2.0