Reconstructing an ancient language is painstaking work. For hundreds of hours, linguists pore over sounds and words in modern languages to create hypotheses of the earliest versions.
Until recently, scientists had not successfully automated the process on a scale of this magnitude. In a paper published in the Proceedings of the National Academies of Sciences this month, four researchers unveiled a computer program to do just this.
Some have described it as a “time machine” that can spirit a linguist thousands of years into the past. Coauthors Alexandre Bouchard-Côté and Tom Griffiths liken their methods to sequencing an ancient organism’s DNA, which reveals essential clues about an earlier habitat and ecosystem.
MIT video analysis experts have developed a new way to amplify subtle shifts in color and motion that are normally invisible to the naked eye.
The result of their work: video that could be used to accurately detect pulse rate based on the human face’s rhythmic flush, monitor babies’ breathing or study the movements of buildings, cranes and mechanical devices.
“You can think about what we’ve made like a microscope, except for video,” says doctoral student Michael Rubinstein, whose team came up with the now patented analytical process they call Eulerian video magnification. “It’s a tool to amplify small spatio-temporal variations you can’t normally see.”
See the video after the jump.
Scientists in fields from astronomy to biochemistry are tapping into the analytical power of groups to complete research that would have been too time consuming to pursue years ago.
By getting a helping hand from legions of online volunteers who work through images and project-related puzzles—an approach called citizen science—researchers are able to tackle and create massive volumes of data. The crowdsourcing effort has already shown dividends in scientific discoveries, published peer-reviewed journal articles, changes in public policies and increasing public interest in the sciences.
“It’s a form of collaboration between professional researchers and the public on outstanding topics of inquiry,” says Andrea Wiggins, a Cornell University and University of New Mexico postdoctoral fellow who studies public participation in data-intensive scientific collaboration. “It’s a phenomenon where machines are helping humans do scientific tasks that the machines can’t yet do well enough because these tasks require human cognition.”
When Dr. Mary Lou Jepsen gets to work on a technology problem, expect big things to happen.
Confronted with the fact that an inaccessible digital world would freeze poor countries out of development, she built the $100 laptop. She designed it, invented or co-invented lots of parts and led a team to bring it into mass production.
The whole project, which has gone on to transform educational opportunities in the developing world, took her three years. Since then, she has continued to make a name for herself as an innovator in holography, display technology and optics.
She earned undergraduate degrees from Brown University in electrical engineering and art, a master’s degree in holography from MIT’s Media Lab and a Ph.D. in Optics from Brown. She has won numerous awards including the 2011 Edwin Land Medal from the Optical Society of America and the 2011 Anita Borg Institute “Women of Vision” Award for Innovation.
Txchnologist: What are you focusing your attention on these days?
Mary Lou Jepsen: I’m now working on next-generation device architectures, which are being designed by thinking first about the screen and how we see light. The device, let’s face it, is increasingly just the screen.