Source: TechCrunch, Dec 2011
Source: TechCrunch, Dec 2011
http://www.reddit.com/r/IAmA/comments/ntsco/i_am_salman_khan_founder_of_khan_academyama/, and his comment about public education policy:
Salman shares his predictions about Education in 2060
Source: MindShift, Nov 2011
We study the influence of television translation techniques on the quality of the English spoken across the EU and OCDE. We identify a large positive effect for subtitled original version as opposed to dubbed television, which loosely corresponds to between four and twenty years of compulsory English education at school.
We also show that the importance of subtitled television is robust to a wide array of specifications. We then fi nd that subtitling and better English skills have an influence on high-tech exports, international student mobility, and other economic and social outcomes.
We therefore provide empirical evidence that, ceteris paribus, English is better in countries where television is in original version with subtitles. The magnitude of the subtitling effect is very large, corresponding to between four and twenty years of English learning at school, and the interaction effects indicate some complementarity between subtitling and formal learning. Pupils in countries where there are subtitles benefit more from their English classes.
The general message in this paper is simple.
Subtitled original version fiction provides continuous exposure to foreign languages. The US is by far the largest producer of fiction programmes shown around the world, so when someone watches a television fi lm in original version, it is very likely that the language source will be English.
Subtitled television programmes then improve the English skills of the viewers, and, thus, the citizens of countries where films are shown in original version speak better English than those where television is dubbed.
We show that dubbing and subtitling countries do not differ signifi cantly in wealth per capita or length of formal English education. Yet there are striking differences in their English skills. Subtitling countries score 77 points higher in the TOEFL, and obtain 23 points more in the EU Survey of English proficiency.
We show in panel regressions that the differences in English skill can be significantly explained by the film translation method used in the country. We identify an effect equivalent to between four and twenty years of English education at school. Our results are robust to the inclusion of other determinants of English skill, like wealth or economic development.
Many of the Khan Academy “practice” exercises have English text; thus the student needs to be reasonably proficient in English.
A motivated student (either self-motivation, or through parents’ urging) can learn to associate the right English terms (narration) with the Malay sub-titles, and gradually learn sufficient English to answer the “practice” exercises.
” …. allow for the individual assessment of any student’s work and allow students who demonstrate their mastery of subjects to earn a certificate of completion awarded byMITx”
Click on the image to view a video of his presentation.
My notes of the presentation:
Source: BlendMyLearning.Com, Dec 2011
This white paper is a comprehensive overview of the Blend My Learning pilot that set out to test blended learning in a real world setting. Written for educators, funders, policy makers, and the general public interested in how technology will transform schools, the paper focuses on the insights and lessons learned from the Khan Academy pilot in Oakland, CA. The paper begins with an executive summary and then discusses:
Source: David Hu blog, Nov 2011
The Khan Academy is well known for its extensive library of over 2600 video lessons. It should also be known for its rapidly-growing set of now 225 exercises — outnumbering stitches on a baseball — with close to 2 million problems done each day.
To determine when a student has finished a certain exercise, we award proficiency to a user who has answered at least 10 problems in a row correctly — known as a streak. Proficiency manifests itself as a gold star, a green patch on teachers’ dashboards, a requirement for some badges (eg. gain 3 proficiencies), and a bounty of “energy” points. Basically, it means we think you’ve mastered the concept and can move on in your quest to know everything.
It turns out that the streak model has serious flaws.
First, if we define proficiency as your chance of getting the next problem correct being above a certain threshold, then the streak becomes a poor binary classifier. Experiments conducted on our data showed a significant difference between students who take, say, 30 problems to get a streak vs. 10 problems right off the bat — the former group was much more likely to miss the next problem after a break than the latter.
False positives is not our only problem, but also false negatives. One of our largest source of complaints is from frustrated students who lost their streak. You get 9 correct, make a silly typo, and lose all your hard-earned progress. In other words, the streak thinks that users who have gotten 9 right and 1 wrong are at the same level as those who haven’t started.
These findings, presented by one of our full-time volunteers Jace, led us to investigate whether we could construct a better proficiency model. We prototyped a constant acceleration “rocketship” model (with heavy gnomes that slow you down on wrong answers), but ultimately decided that a prudent first step would be to just abstract away the streak model with the notion of “fill up the bar”.
Conversations with the team led me to conceive of applying machine learning to predict the likelihood of getting the next problem correct, and use that as the basis for a new proficiency model. Basically, if we think you’re more than % likely to get the next problem correct, for some threshold , we’ll say you’re proficient.
Source: Inside Higher Ed, Dec 2011
The real revolution at Khan Academy, they contend, is not being streamed on its vaunted website; it is happening in the back end of the platform, where Khan’s engineers are learning as much about the site’s 1.4 million registered users as those users are learning about math and science.
Using math and computer science concepts decidedly more advanced than most of those in Khan’s video library, the Khan engineers have trained the website’s exercise platform how to predict, with startling accuracy, how likely it is that a student will correctly answer the next practice problem — and whether that student will be able to solve the same type of problem a week, two weeks, and a month later.
They do this by accounting for hundreds of data points that describe, in numbers, the entire history of the relationship between a learner and a concept.