Microsoft research has recently found a solution that will make photographer’s work easier. A new technology has been developed by Microsoft Corporation, which is said to boost artificial intelligence studies for the company. The technology will be beneficial for a lot of people because it uses a technical neurological network that is similar to the functional design of the human brain. The technology is made to caption a photo and place it in its correct category, by linking vision to language, and creating better methods and results that researchers have been trying to master for years.
The advances of the technology began last year. However, in a recent announcement released May 29, from Redmond-based company explained how the new technology system will work. Researcher, Xioadong He, of Microsoft Research’s Deep Learning Technology Center told sources, the machine was trained to understand how humans understand images. The machine can speak, understand things, hear, and see the same way a human does. In the blog, the technology is said to observe the way humans use a computer and it has the capability to decide to focus on the human and not the background.
Microsoft Corporation boosts artificial intelligence and technology because the creation uses features such as facial hair to identify the gender of the human it is focusing on. Microsoft Research’s Natural Language Processing Group, Margaret Mitchell, told sources, “the technology has gone from not working at all to working, due to the neural network.” Although the technology is not yet perfect, the captioning photos concept is what makes the Microsoft Corporation technology catchy.
Microsoft has come up with an algorithm which breaks down the steps of how the technology captions an image. The systems starts with a convolutional neural network, that identifies words that describes the image. Since the neural network has been fed a number of captions and images it can easily draw from the creative base to determine what the object, person, or scenery is. Microsoft Corporation trained the technology in comparing shapes, color palettes, and it goes into its built in language model which helps it grammatically correct captions and arrange words into sentences. Once the caption is formed, a checker is posted to make sure an accurate caption correlates with the image.
Researchers say the technology continues to improve, and they see the extensive amount of possibilities of how the tools can be used for important gains for artificial intelligence, since personal computers are becoming more intelligent. He, told sources the technology could be a beneficial element by providing a more sophisticated tool for robot Al. He believes the Microsoft Corporation technology is a universal tool that helps individuals get a better understanding of the world around them, and it helps when a person needs a better understanding of what is in front of them whether it is a picture that has recently been taken, or an image from the past. He also stated, since techies have been looking forward to such a creation, having the technology turns the near future into an optimistic one because the only thing left to do is improve the technology now that a solution has been found.
Microsoft Corporation boosts of artificial intelligence and technology is said to be one of the best researches for automated captioning of photos in a very long time. When the technology was referenced at the latest competition MS COCO Captioning Challenge of 2015, university experts and corporate researchers have begun to compete against one another in creating the best algorithm that offers an easy explanation of the automated image captioning device. MS COCO Captioning Challenge ends on Friday, and a leaderboard has been tracking how well each team is doing. Judges are said to announce the top performers at the CVR computer vision conference early June.
By Krystle Mitchell
Bidness etc: Microsoft Corporation Advances Artificial Intelligence
Softpedia: Microsoft creates software that detects everything in photos
Microsoftblog: Microsoft Research project interprets caption photos
Photo Courtesy of Michael Kappel’s Flickr Page -Creative Commons License