In the recent years there has been a huge leap in the field of artificial intelligence. Although we have not been able to create a fully conscious artificial intelligence, but what we have been able to create are bits and pieces of the puzzle.
Now what lies ahead of us is to wait and see when and how the puzzle is being solved. Recently we saw an interesting demonstration of Microsoft image recognition API at how-old.net, which showed that machines are capable of identifying the age and gender of a person through a photograph.
In another such advancement, Stephen wolfram, the creator of siri’s knowledge base has developed a new technology that can identify the subject of an image. Known as the Image identification project, the new tool is one of the many artificial intelligence related tools that wolfram has been developing.
I’ve tried out the image identification tool developed by Wolfram and to be honest, the tool is not always accurate. The tool is good at identifying natural things, such as animals. But when it makes a mistake, it’s just like a human. For example, I uploaded a picture of a street dog and the tool identified it as dog from a different breed. When it comes to identify man made objects, the tool fails miserably.
As of now the tool is capable of identifying more than 10000 common objects and in order to be able to do so, Wolfram had to feed millions of photos into the tool. The tool is expected to learn more as people start uploading more photos to it.