Microsoft Research builds a Drawing Bot which brings your Description to Life
First, it’s not a final product, but a result some interesting engineering from Microsoft Research Team. They had been working on a new AI system which can draw using a technology named as Generative Adversarial Network (GAN). In simple English means a network of bots which decide among themselves to generate the right character based on the description.
So as expected, this AI has two parts, one that deals with the logic of what should be drawn, and second actually draws it. So it’s not just a simple image search made on bing, but an actual picture is drawn pixel by pixel. According to the MS Research Team:
The new artificial intelligence technology under development in Microsoft’s research labs is programmed to pay close attention to individual words when generating images from caption-like text descriptions. This deliberate focus produced a nearly three-fold boost in image quality compared to the previous state-of-the-art technique for text-to-image generation, according to results on an industry standard test reported in a research paper posted on arXiv.org.
To further improve, Research team is trying to put a common sense as well into the AI. This gets accumulated as it gets more data. It starts from one single input, but if the context grows more than that, it starts learning to find what more can be done.
The model also learns what humans call commonsense from the training data, and it pulls on this learned notion to fill in details of images that are left to the imagination. For example, since many images of birds in the training data show birds sitting on tree branches, the AttnGAN usually draws birds sitting on branches unless the text specifies otherwise.
This definitely has a huge scope. I can only imagine my son describing an event that happened to him which becomes an actual picture. Imagine this being used in Police Department where the sketch can be drawn as the description of the suspect is talked about.
