Navigating the future
If you have ever been curious about what the future will look like, take a look at the TED lecture series at www.ted.org The ideas that are being developed are beyond imagination, and it shapes a future unlike anything we could have predicted.
Take the Seadragon project for example :
Under development from Microsoft, Seadragon was designed simply as a new way to look at electronic images and image data. The software was designed to tackle the baffling task of navigating incredible amounts of image data quickly and efficiently. It may not sound like much of an accomplishment, but imagine storing entire books of photographs, with text and magazines on your mobile phone. Imagine browsing through the data effortlessly on your Iphone. Imagine it being easier to navigate and more convenient than a good old fashioned magazine. It is actually quite daunting, and despite our best efforts with technology, nothing has come close to replacing printed media.
But in a sense, replacing it isn’t what this software intends to do. By completely rethinking the way we navigate visual media, there may be an opportunity to improve on it. As with any technological leap, it is only when something better comes along that we willingly shift our preferences to embrace something new.
But it didn’t stop with navigating visual media. Seadragon had developed sophisticated algorithms for managing image data, and one aspect of its functionality was to sort and match similar images. Imagine, for example, if you took a picture of a cookie. Imagine if you took lots of pictures of lots of cookies. Maybe you took lots of pictures of potatoes and trees as well. Now imagine if you could use your computer to search for ‘cookie’ and find only images of cookies. Sounds easy, right?
Well not really.
Up until now, it takes a human brain to associate the image of a cookie with the word ‘cookie’. Even Google images relies on typed captions under photographs to provide a description; it really can’t tell what a photograph is unless some human brain recognized the image and wrote caption for it. Our brains, extraordinary as they are, are capable of matching patterns and making associations that computers struggle with, and even a database as powerful as Google needs a lot of human brains to sort the images.
So here is where Seadragon accomplished something extraordinary. Seadragon developed the capacity to make sense of all of these images. The software was able to recognize some of the key features of image data, to recognize patterns and make associations with other images. The software couldn’t recognize an image of a cookie, per se, but it could compare two images that looked similar, and make a connection. It could scan thousands of images, and make associations that previously took a human brain. So now one human brain labels one cookie, and the software can make the connection to thousands of images of cookies.
Now apply this to the entire web. Imagine the vast amounts of image data available online; photographs, drawings, maps, paintings. The truth is, only a small percentage of the image data currently available on the web is searchable, because up until now, the only way to search anything with a computer was to rely on text descriptions. But imagine if we could use images as a search tool? Imagine browsing for long lost friends from a photograph. Imagine sketching the shape of a particular flower and finding it instantly over the internet. Imagine a completely new way of navigating data that no longer relies on a keyboard; a fundamental shift in how we index and navigate data.
But why stop with images? What about sounds? Music? Enter your favorite song at www.pandora.com and it generates a non-stop stream of music from similar artists; music you might like. It does this by gathering input from all of its users. Don’t like a suggested song? Click on the website to skip it. The software depends on this user input to make associations between music and preferences; essentially associating musical preferences based on the collective input of thousands of users. Pandora is one of the new developments in interactive internet that gather input from its users, essentially gathering intelligence and becoming more effective and efficient. An extremely powerful approach to organizing data that relies on the most powerful computer of all.
The human brain.
Daniel Ferguson
Uncategorized | admin | 
March 31, 2009 3:02 pm
By ben, April 26, 2009 @ 9:45 am
I absolutely LOVE the TED conference videos. I could spend all day watching those and frequently have to hold back from publishing too many links to it on my own blog. I recently watched the one about predicting political situations with computer models… using the example of Iran and nuclear weapons. So very cool.