2014/06/30 § 1 Comment
지난주에 김기응 교수님과 식사하며 이런 저런 이야기를 나누었다. 교수님 랩에서는 이번 여름방학에 Bishop 책으로 기계학습 세미나를 할 계획이라고 하신다. 목표는 이 책을 모두 끝내는 것. 기계학습을 하려면 이 책은 기본으로 읽어봐야 하지 않겠냐고 하셨다. 미국 나가기 전에 이 책을 다 읽고 가라고 하시면서, 반드시 연습문제를 풀어보라고 하셨다. 나중에는 공부할 시간이 부족하기 때문에, 결국 지금 공부해 놓은 것으로 앞으로 먹고 사는 것이라고.
만약 교육 관련된 연구를 하게 된다면, CS 학생으로서 내 포지션이 애매하다고 느꼈었다. 교육 전공자도 아니고 순수 컴퓨터 연구도 아닌 것이… 교수님과 이야기하고 나니, 이 책을 읽는 것이 내 포지션에서 해야할 일이라는 생각이 들었다. 여름 방학에 뭘 해야 후회하지 않을까 고민중이었는데, 쉬면서 연습문제 풀어보는 것도 나쁘지 않을 것 같다. 음..ㅋㅋㅋㅋㅋ
[Scrap] Google Sponsors Carnegie Mellon Research To Improve Effectiveness of Online Education — Paying Attention To How People Learn Promises To Enhance MOOCs
2014/06/25 § Leave a comment
PITTSBURGH—Carnegie Mellon University researchers will tap data-driven approaches to improving learning as part of a new Google-sponsored effort to unlock the educational potential of massive open online courses, or MOOCs.
The multi-year program, made possible through a Google Focused Research Award, employs a variety of methods to improve MOOCs. The research plan includes development of techniques for automatically analyzing and providing feedback on student work, for creating social ties between learners and for designing MOOCS that are effective for students with a variety of cultural backgrounds.
“We’re excited about Carnegie Mellon’s work on mechanisms to allow online courses to adapt automatically to the learning needs of individual students,” says Alfred Spector, vice president of Research and Special Initiatives at Google. “We believe this research will make online courses much more engaging, and benefit both students and educators around the world.”
The goal is to get online courses to be as successful as the best courses in brick-and-mortar classrooms, said Justine Cassell, associate vice provost of technology strategy and impact and co-director of CMU’s Simon Initiative, a university-wide effort that uses learning science and technology to improve student learning.
“A MOOC today typically means a lecture-style presentation with little if any opportunity for interaction with other people in the course,” Cassell said. Not surprisingly, most students drop out long before the courses are complete, and learning gains are often low even for those who stick it out. “Unless the MOOCs pay attention to how people actually learn, they will not be able to improve effectiveness, and will end up as just a passing fad,” she added.
The CMU program will approach the problem from several directions.
In one thrust, Emma Brunskill, assistant professor of computer science, and Ken Koedinger, professor in theHuman-Computer Interaction Institute (HCII) and director of the Pittsburgh Science of Learning Center, will use machine-learning techniques to personalize the MOOC learning experience. Computer programs will evaluate each student’s work, identifying subject matter that has been mastered and areas where additional study or different types of exercises could be beneficial.
This kind of data-driven learning, pioneered at Carnegie Mellon, has shown to make learning faster and more effective.
In another thrust Carolyn Rosé, associate professor in the Language Technologies Institute, and Robert Kraut, professor of HCII, will look for ways to reduce attrition. Though many students in MOOCs are simply browsing, even committed students have a high drop-out rate. To improve retention, Kraut and Rosé will look for ways of increasing socialization, through mentoring, team assignments and other techniques. They also seek to identify warning signs that students are in danger of dropping out and to develop interventions to re-engage them in courses.
A third thrust will focus on how to make the content of the course more engaging. Jessica Hammer, assistant professor in HCII and the Entertainment Technology Center, and Amy Ogan, assistant professor in HCII, will examine how to enhance the pleasurable aspects of MOOCs, adjust the design of courses available globally to account for cultural differences, and develop a deeper understanding of how and when to incorporate game play into MOOCs.
The Google award will fund the research at $300,000 a year for two years, with an option for a third year.
Google’s Spector is a member of the Global Learning Council (GLC), which includes leaders from academia, the private sector and the foundation community. Chaired by Carnegie Mellon President Subra Suresh, the GLC aims to serve as a standards and best-practices resource for individuals and organizations seeking to deploy technology-enhanced learning approaches to improve learning outcomes. The GLC will have its first annual meeting in Pittsburgh in September.
The Human-Computer Interaction Institute, Language Technologies Institute and Computer Science Department are all part of Carnegie Mellon’s top-ranked School of Computer Science, which is celebrating its 25th year. Follow the school on Twitter @SCSatCMU.
2014/06/25 § Leave a comment
2014/06/09 § Leave a comment
2014/06/07 § Leave a comment
A good reason to draw letters when memorizing English words? Seeing is not enough:
In another study, Dr. James is comparing children who physically form letters with those who only watch others doing it. Her observations suggest that it is only the actual effort that engages the brain’s motor pathways and delivers the learning benefits of handwriting.
Handwriting is important (compared with writing on computers):
A 2012 study led by Karin James, a psychologist at Indiana University, lent support to that view.
Children who had not yet learned to read and write were presented with a letter or a shape on an index card and asked to reproduce it in one of three ways: trace the image on a page with a dotted outline, draw it on a blank white sheet, or type it on a computer. They were then placed in a brain scanner and shown the image again.
The researchers found that the initial duplication process mattered a great deal. When children had drawn a letter freehand, they exhibited increased activity in three areas of the brain that are activated in adults when they read and write: the left fusiform gyrus, the inferior frontal gyrus and the posterior parietal cortex.
By contrast, children who typed or traced the letter or shape showed no such effect. The activation was significantly weaker.