By Turing Robot, Special for USDR
CES 2016 (2016 International Consumer Electronics Show) is being held on January 6, 2016 in Las Vegas. At CES 2016, robots remain one of the hottest topics. Turing Robot brings out a unified hardware and software product solution plan via Turing OS.
Turing OS makes it possible for robots to precisely receive and understand external information such as voices, visuals, texts and body movements delivered in multimode like humans. It also enables robots to express variable information, perform multimode reaction to external information received and achieve humanoid communication ability.
Thinking ability is often viewed as a way to judge the intelligence level of a person, which is combined by all kinds of thinking modes. With the progression of human recognition ability, the amounts of thinking modes increase dramatically, and the ability of human thoughts has increased as well.
Research shows that humans can conduct effective training on their thinking modes in variable ways, including reverse thinking, deductive thinking and so on. As a result, for robots that keep “pursuing human intelligence”, it is especially vital to learn the humanoid thinking mode or the mechanism of thinking ability.
Turing OS not only can recognize positive and negative emotions of human beings, but can also judge from the dominant and recessive factors in those emotions. This means that robots can recognize whether its communicator is glad or sad, and at the same time, decide the degrees of his happiness and sorrow based on the effective information the communicator expressed. As for the emotional expression, combined with the multimode interaction method of Turing Robot, robots can output emotions multi-dimensionally from voice, facial expressions or body movements.
The Affective Computing Engine of Turing Robot supports 25 kinds of linguistic emotion recognition currently with an accuracy of 95.1%. In emotional expression area, Turing OS boasts 468 kinds of emotional voice expression, 88 sets of facial-movement expression and 120 kinds of voice tones, which equip robots with extreme closeness to humanoid emotion ability.
To ensure the continuous increase of the intelligence level of robots and make it possible for robots to cognitively learn like humans, Turing OS includes a Self-Learning Engine based on a D-RNN algorithm. Combined with big data contributed by 130,000 partners gathered on the Turing Robot open platform and the world’s fastest super computer “TH-2” to do the computing, the self-learning iteration speed of Turing OS has jumped from 7 days to online real-time processing.
SOURCE Turing Robot