October 29, 2010
Citizen Scientists Explore Ancient Mongolia from Afar
"Citizen archaeologists" helped researchers find Bronze Age burial sites and other Mongolian antiquities as part of a new National Geographic-supported expedition this summer. The groundbreaking "Field Expedition: Mongolia "” Valley of the Khans Project" invited Web users around the world to join a field expedition online in real time as "citizen scientists" from the comfort of their homes. The online expedition continues at http://exploration.nationalgeographic.com/mongolia.
The field expedition, headed by National Geographic Emerging Explorer Albert Yu-Min Lin of the University of California, San Diego, in collaboration with professors Shagdaryn Bira and Tsogt-Ochiryn Ishdorj of the International Association for Mongol Studies, and National Geographic Archaeology Fellow Fredrik Hiebert, is using modern, noninvasive tools to explore and map parts of Mongolia including the "forbidden precinct" "” the homeland of Genghis Khan, which has gone unexplored for 800 years. The project is searching for archaeological sites that shed light on the country's rich cultural history and heritage while maintaining respect for local customs and beliefs.
In a first for National Geographic Digital Media, "Field Expedition: Mongolia" features an online component that encourages active participation by the public "” known as human computation, or "crowdsourcing" "” through a customized interactive Web portal. Using map-accurate, high-resolution satellite imagery made available by the GeoEye Foundation, citizen scientists analyzed real-time data, maps and other information direct from the field to mark anomalies that might represent archaeological ruins. The field team then used information the taggers collected to "ground-truth" the tagged sites, employing ground-penetrating radar, unmanned aerial vehicles, remote sensors and on-site digital archaeology. Human computation is an ongoing research focus at UCSD's California Institute for Telecommunications and Information Technology (Calit2).
"The power of crowdsourcing is utilizing the very human traits that we take for granted, such as the fact that we can identify something that is natural versus something that is man-made," explains Lin, who is affiliated with Calit2's Center of Interdisciplinary Science for Art, Architecture and Archaeology (CISA3). "It's hard to write a computer program that will tell you the same thing."
Still ongoing, "Field Expedition: Mongolia" launched June 9, 2010, and has drawn more than 6,000 taggers who have pinpointed more than a million specific areas of interest. One of the most active taggers has been Allison Shefcyk, a 24-year-old from West Hartford, Conn., who has yearned to be an archaeologist since childhood. Shefcyk said she knew little about Mongolia before embarking on the search. "When I first heard about the expedition, I decided I had to learn more about Mongolia in order to get a better idea of what I was looking at," she said. "That meant looking carefully at the blog (on the National Geographic website), going to the library, etc. I quickly became captivated by this unique and beautiful country." Shefcyk has tagged about 50,000 sites.
Satellite imagery taggers identified a wide range of interest points on the ground in the vast open spaces of Mongolia. The team rushed to one promising site only to find it was actually a herd of sheep. However, other leads were more fruitful. Some were found to be "khirigsurs," large tombs encircled by rings of rocks that date to the Bronze Age, more than 3,000 years old. Some sites were found to be looted, perhaps recently. Team archaeologists are analyzing the data collected.
Archaeology in Mongolia is complicated by the nomadic nature of its historic inhabitants and the size of the country "” Mongolia is slightly smaller than the state of Alaska. "Perhaps the main goal of the Valley of the Khans project is to help give these people a voice and to protect this area that abounds with ancient history," the expedition blog states. "By learning about the history of this region, we can prove its significance to Mongolian life and a heritage that shaped the world."
The hope is that many of those landmarks will eventually be protected as UNESCO World Heritage Sites. "It's almost like the crowd is guiding us," Lin says. "We're entering a new information processing era where the networks being created through the Internet are allowing us to approach collaboration on a whole new scale."
High-resolution satellite imagery for "Field Expedition: Mongolia" is provided by the GeoEye Foundation, a nonprofit organization established in 2007 to help train others to map, monitor and measure the Earth. The Foundation's focus is to foster the growth of the next generation of geospatial technology professionals; provide satellite imagery to students and faculty to advance research in geographic information systems and technology as well as environmental studies; and assist nongovernmental organizations in humanitarian support missions. The Virginia-based GeoEye Foundation donated all of the satellite imagery and arranged for additional satellite "fly-overs" when new images over specific areas were needed by the researchers.
A team of computer, electrical and cognitive scientists at UCSD's Calit2 wrote the crowdsourcing algorithm that allowed "Field Expedition: Mongolia" to capture user tags and the back-end scripts that cut GeoEye's satellite images into manageable sizes for users. UCSD scientists wrote the program that provides taggers with feedback on how well they tagged an image, offering advice and showing where others tagged items. UCSD researchers also provided blog updates and reports from the field. These mechanisms combined to allow thousands of online participants to virtually join the expedition team in Mongolia from their homes. Human intuition data points generated by "Field Expedition: Mongolia" taggers help back-end computers learn to accurately identify ancient structures and supplement existing UCSD computer object-recognition algorithms.
The enthusiasm for learning and exploration evidenced during the "Field Expedition: Mongolia" holds great promise for future explorations and research projects in a variety of fields, says Calit2 research scientist Lin. "We're still very far away from the day when we have made human observation obsolete," he says. "Lots of problems in science are computationally exhaustive and most are still beyond the capacity of computing systems. Now you can imagine the next level: Taking that collective thinking and having it guide itself toward the evolution of its own platform, where the crowd determines where it wants to go and what kinds of problems it wants to solve."
Digitaria, a leading digital marketing and technology firm, designed the look and feel of the "Field Expedition: Mongolia" website, and developed the flash-driven front-end interface that displays the satellite images and tools used to tag the features, including roads, rivers and ancient structures. The interface was created to meet the design criteria of UCSD scientists, whose goal is developing intuitive mechanisms and motivational elements that allow the crowd to be harnessed as a data-gathering machine. This process, human computation, enables large numbers of people who lack relevant training to help the scientists accurately analyze huge volumes of data. Digitaria also built out the registration process and blog template.
National Geographic Digital Media (NGDM) is the multimedia division of National Geographic Ventures, the wholly owned, taxable subsidiary of the National Geographic Society, one of the world's largest educational and scientific nonprofit organizations, working to inspire people to care about the planet. Holding many top industry awards, NGDM publishes www.nationalgeographic.com; produces short-form video for broadband markets; manages marketing and content partnerships across broadband, mobile, gaming and other consumer digital platforms; and provides video and film footage to commercial, theatrical, educational and other digital footage markets.
Image Caption: A team of computer, electrical and cognitive scientists at UCSD's Calit2 wrote the crowdsourcing algorithm that allowed "Field Expedition: Mongolia" to capture user tags and the back-end scripts that cut GeoEye's satellite images into manageable sizes for users.
On the Net: