Machine Learning Algorithms require a lot of training data. For example, say you want to create a program separating pictures of cars by manufacturer. To train the algorithm, you might need a few thousand pictures that are classified as Hondas, Toyotas, etc. This is easy but cumbersome for humans to do.
The point is, we use Amazon Mechanical Turk to create training sets for Machine Learning algorithms, which I always find really interesting and clever.
"Another cool creation adding human intelligence to software is VizWiz which aids vision-impaired individuals answer visual environment oriented questions such as who sent you the letter, and distinguish various items in your pantry. Some of these questions are answered as quick as 59 seconds. " - Taken from http://www.trendecommerce.com/experiment-post-amazon-mechani...
I experimented with having turkers grade elementary school math quizzes. The idea was that teachers would scan in their all their quizzes/tests/homework to be graded along with the answer key, and turkers would grade according to the key. It worked, but the margins weren't there.
8 comments
[ 3.3 ms ] story [ 28.1 ms ] threadhttp://www.tenthousandcents.com/top.html
The point is, we use Amazon Mechanical Turk to create training sets for Machine Learning algorithms, which I always find really interesting and clever.
http://www.wired.com/techbiz/it/news/2007/11/fossett_search
This was more of a proof of concept but was pretty impressive. http://code.recollect.com/post/13372390010/mechanical-turk-u...
"Another cool creation adding human intelligence to software is VizWiz which aids vision-impaired individuals answer visual environment oriented questions such as who sent you the letter, and distinguish various items in your pantry. Some of these questions are answered as quick as 59 seconds. " - Taken from http://www.trendecommerce.com/experiment-post-amazon-mechani...