Just wanted to draw attention to a lovely lecture by Luis Von Ahn from a couple of years ago.
He takes us through a particularly entertaining look at what humans can do better than computers, and ironically, how to use computers to make use of this resource.
He himself has devised a game that allows players, unbenknownst to themselves, to usefully describe and tag images from the internet. Two random players somewhere in the world are presented with exactly the same image at the same time - the idea of the game is for the two users is to type in the same word to sum up the image in as few trys as possible, without any other way of communicating with each other. When the two players reach the same word they receive a score based on how long it took them to come to agree. Of course in attempting to arrive at the same word they build up a very useful set of descriptors around the image they are looking at.
The image tagging by-product of the game could, if used in conjunction with something like Google Image Labeler, swiftly provide rich textual descriptions for most of the images on the internet and therefore much improve visually impaired people's experience of using the web pretty much without anyone actually working to make it happen.
RSS Feed