Google’s DeepMind team created a human-like robot that can mimic our speech sound. The Artificial Intelligence robot makes use of neuroscience and deep learning automation. The robot can speak with such an elaborate fluency that it can fool anybody into thinking they’re speaking with a real human. That typical robot voice might be a thing of the past now.
DeepMind is a technology which has been developed by an English company. Google saw the potential and decided to buy the tech startup in 2014. The team at DeepMind aims to create such advanced computers that they can effortlessly act as humans. The developers have made the system quite different from other similar AI technologies such as DeepBlue and Watson. DeepMind is made to perform any task necessary without having to be programmed in advance.
The program, called WaveNet, hopes to make its machinery speak in a fluent human voice. Those of you that have watched the movie “Her” may understand the concept better.
“WaveNets are able to generate speech which mimics any human voice and which sounds more natural than the best existing Text-to-Speech systems, reducing the gap with human performance by over 50%,” says the DeepMind team in a recent blog post.
Yet how keen are we to hold conversations with machines? The scientists at DeepMind consider it a “long-standing dream”. This is why Google has developed various networks over the year to increase the computers’ ability to understand the human speech.
Artificial speech programs are quite difficult to create. There are only two methods currently used for this purpose. The first one involves collecting an extremely large batch of words or fragments. The second one takes a more artificial approach. Words are formed electronically, and they are easier to tweak compared to the first process. However, the resulting sound is too “robotic”.
DeepMind decided to ditch the existing technologies and create WaveNets. The technology can also be used to generate music. This opens new doors to audio modelling and music generation. The team’s paper, called “WaveNet: A Generative Model for Raw Audio”, can be read here.
Although Google’s DeepMind team created a human-like robot, the technology is still in its beta phase. And we are probably going to have to wait a while until actually seeing the technology integrated into a Google device.
IMAGE SOURCE: ILikeIt