Data Innovators Rinat Sergeev, Data Scientist at NASA Tournament Lab

Published on September 6th, 2013 | by Travis Korte

0

5 Questions for Crowdsourcing Expert Rinat Sergeev

The Center for Data Innovation spoke with the NASA Tournament Lab‘s (NTL) Data Science lead, Rinat Sergeev, about the philosophy of crowdsourcing and its place in government. Sergeev, a physicist by training, studies crowdsourcing from a methodological perspective, and works with the rest of the NTL team to facilitate public challenges to solve complex problems NASA faces.

Travis Korte: Tell me a little bit about the vision for NTL? What is its purpose, in your view?

Rinat Sergeev: The foundation of our lab reflects the growing value of crowdsourcing that’s finally been recognized by major US government agencies such as NASA. The government considers crowdsourcing an extremely promising and efficient way to introduce innovations, reduce their overhead, and cut through the inefficiency bottlenecks that all big organizations have. The purpose of our lab is to cover both operational and explorational approaches to crowdsourcing. We build up and use our expertise by running hands-on projects from NASA and collaborating with agencies on professional crowdsourcing platforms (primarily on TopCoder, which has a community of over 500,000 people with an abundance of algorithmic, software, and computer design skills).

At the same time, we study the ups and downs of crowdsourcing as a tool, learning when and how to use it most efficiently and providing academic expertise both to crowdsourcing platforms (how to set the contests and incentives correctly) and to NASA/agencies regarding how to adopt crowdsourcing mechanisms and integrate them into their internal workflows.

TK: What are some of the sorts of problems that NASA has looked to the data science community for in the past?

RS: One of the advantages of crowdsourcing is its incredible flexibility–there’s a large variety of problem types you can solve through it. The problems can range from small graphic or ideation contests to software apps to large data-intensive software and algorithmic contests.

Among the completed projects, the most data-intensive one was to solve the problem of International Space Station solar power supply (the linked video has also been created via crowdsourcing).

The project allowed us to combine multiple strong sides of the community in order to develop the best algorithm for Space Station solar panels, one that ensures they shadow each other as little as possible. The community was able to recreate and adopt for the final contest the whole 3D model of the Space Station, proving that there is no limit on the complexity of a problem for crowdsourcing.

Currently, we have three heavy-data projects on the latest stage of preparation, covering image-recognition, time-series pattern-recognition and prediction, and machine-learning techniques.

TK: What, in your mind, is the purpose of these hackathons and competitions?

RS: Basically, it’s a completely new form of work organization, which became possible only after the boom of social networks, that allows you to organize your workflow around those incentives that are strongly undervalued in traditional business models.

As a result, it gives the opportunity to overcome traditional models’ inefficiencies, making crowdsourcing an extremely efficient and complimentary option to traditional models.

TK: What can government organizations do to better engage and work with the data science community?

RS: The recognition of the importance of professional communities is on the rise in government organizations. At the same time, government organizations are traditionally inflexible in introducing new, non-conventional forms of cooperation, and professional communities, on the other hand, are rarely providing the organizations with bureaucratically recognizable tools of engagement. In those cases when the effort is made from the both sides (NASA establishing NTL to understand crowdsourcing, and TopCoder developing and promoting a very formal, structured, verifiable, and quantifiable approach to community), the engagement works.

TK: Talk a little about how some of the challenges associated with actually implementing projects sourced from programming competitions and hackathons, and what NTL has done to address these challenges.

RS: When you are exploring completely new grounds, every step you make is a challenge: from initial engagement on the project to final evaluation, implementation, and support of the results. We have learned a lot and we are still learning. One of the hardest challenges we have to deal with is a huge difference in flexibility and pace between the organizations and communities. Combining and managing those within one big project looks a lot like making a square canister fit into a round hole, like they did during the Apollo 13 mission. We found that solving the problems of organizations via crowdsourcing is actually the best way to learn something not only about crowdsourcing, but organizations as well.

Tags: , , , , , ,


About the Author

Travis Korte is a research analyst at the Center for Data Innovation specializing in data science applications and open data. He has a background in journalism, computer science and statistics. Prior to joining the Center for Data Innovation, he launched the Science vertical of The Huffington Post and served as its Associate Editor, covering a wide range of science and technology topics. He has worked on data science projects with HuffPost and other organizations. Before this, he graduated with highest honors from the University of California, Berkeley, having studied critical theory and completed coursework in computer science and economics. His research interests are in computational social science and using data to engage with complex social systems. You can follow him on Twitter @traviskorte.



Back to Top ↑

Show Buttons
Hide Buttons