IDG (International Data Group) predicts the use of unstructured Big Data will steadily grow at a rate of 62 percent annually. They also predict by the year 2022, 93 percent of all data being processed will be unstructured. The issue of processing unstructured Big Data is currently one of the most challenging problems facing Data Scientists, and Spare5 has met the problem head on it with their Intelligent Crowdsourcing solution. The clients who have signed on with Spare5 (and have various use cases available for perusal) include Getty Images, Avvo, GoPro, IBM Expedia, Sentient Technologies, and many others.
Spare5 Provides Solutions
Spare5 is a pioneering company. Its staff is creative and focused on being flexible and adaptable, as they look for new applications for their services. They have created a unique program, in that it combines Machine Learning with human insights. Spare5 uses a community of experts and specialists to perform micro-tasks tailored to the client’s needs. After being screened for quality, Spare5 clients can use their platform to train Artificial Intelligence (AI) systems, improve on their browse-and-search experiences, enhance their directories, and generally work more efficiently.
Myles Brundage, the Director at Sentient Technologies, said:
“Spare5 is a valuable part of our product development for Sentient Aware, our AI-powered shopping assistant. With Spare5’s unique ability to access people with specific domain experience, we are able to quickly validate our AI-generated models by comparing them to how people perceive certain nuances between different retail products.”
Spare5’s platform uses a variation of the crowdsourcing technique, which they call “Intelligent Crowdsourcing.” Intelligent Crowdsourcing uses a network of qualified individuals as their “crowd,” while traditional crowdsourcing uses its customer base or the general population. This model allows Spare5 to engage the right person quickly and efficiently, as they deliver insights on unstructured data. Spare5 has built a community of specialists (called “Fives” within the organization) with a wide range of skills and interests, and has also developed a library of pre-screening, game-like tasks. The pre-screening process assures the right person is assigned the right task. This approach provides the best micro-task advice available from the community.
Andy Ganse, the Principal Data Scientist at Spare5, explained:
“A micro-task is a bite-size question, or challenge, that usually takes anywhere from a few seconds to a few minutes. Part of what Spare5 specializes in, is taking these enormous data challenges and sticking them into tiny little tasks that people can complete in their free time. For example, waiting for an airplane, or for a coffee, or while they commute on a bus. We call them micro-tasks because they’re so small, and we do our best to create a game-like experience.”
A detailed quality assurance process then filters the task results for accuracy. The process includes Spare5’s Machine Learning algorithms. Machine Learning takes place while the client uses their platform, making the process becomes faster, better, and smarter. By combining human insights and Machine Learning, customers receive clean, labeled, in-house data that otherwise would remain unused and unidentified. A variety of APIs and SDKs process the data and integrate it into existing data workflows, which can then be used to study customer bases and produce highly functional business reports.
The Spare5 Mission
Matt Bencke, co-founder and CEO of Spare5, said:
“Our mission is to tap the world’s potential brainpower. Businesses need specialized human insights to solve complex data problems. It used to be somewhere between impossible and impractical to crowdsource specialized insights at scale, with confidence, and speed. Not anymore. We love seeing our customers get the help they need to interpret unstructured data, while freeing up their employees to focus on their core competencies. There is a profound difference when the right human intelligence powers Machine Learning. And we’ve just begun.”
Choosing the right person for the right loop is an important aspect of the Spare5. It is generally understood the current state of Machine Learning requires a human be kept in the loop when an algorithm is being trained. A human teacher provides training data and routes tasks when the algorithm does not have a clearly learned path. Spare5 has developed a community of knowledgeable and experienced people who have a proven track record as solid, reliable performers in data tasks.
But Spare5 takes it one step further. They also match the right person to the right loop, and “the task type.” This additional layer makes a difference in quality. It makes sense to have fashionistas deal with fashion tasks, architects dealing with architectural tasks, and writers handling writing tasks. These classifications are very important, and are a major part of what separates Spare5 from a more traditional crowdsourcing format. One of Spare5’s strengths is its ability to choose the right person for the right loop, and then optimizing that choice with their Machine Learning algorithms and workflows.
A 2015 study reported that over 160 million Americans are on their smartphones for nearly 5 hours every day. Spare5 allows businesses to use this kind of unstructured Big Data in ways impossible for individual computers to handle on their own. Tasks such as keyword creation, image tagging, and classifying unstructured data, each require the intelligence of a human, and Spare5 provides this in a reliable, scalable, cost-effective process.
Focusing On Business
Jon Sakoda, a General Partner at NEA remarked:
“Spare5 is strongly focused on business data needs and has built a uniquely engaged and capable community to address those needs with high-quality insights. Their technology platform takes the complexity out of micro-tasking and makes it more rewarding and productive for both customers and contributors. This investment will help them further tap the potential of this vast resource while scaling to meet growing demand.”
Spare5 uses a subscription format, and companies can join the platform with relative ease, in turn gaining access to a range of task templates and a dedicated team member. The subscription price varies with the amount of customization, the complexity of the task, and the specialty.
Spare5 customers agree the Intelligent Crowdsourcing platform reduces their costs and makes the processing of complex data tasks easier. Their platform allows clients to upgrade the quality of their Metadata without spending valuable resources on outsourcing and training. While a large portion of their sales are driven by SEO, and by being easy to find, word of mouth advertising and social media has also played a role in their sales. Spare5 has found a way to successfully integrate the human mind with the training computers, producing some remarkably efficient results. That’s information worth sharing with friends.