With the rapid development of speech recognition, can robots really understand us?

If the machine can understand people, it must go through a training process similar to that of people. People first find phonemes or word syllables representing speech, as well as their frequency, amplitude and time-varying characteristics, and store them in the machine. When the robot recognizes human voice, it first compares the voice parameters to be recognized with the internal parameters of the machine one by one, finds out the most similar reference sound, and outputs the recognition results through display and sound. If there is no similar sound, it will refuse to recognize it. After all, robots can't compare with human intelligence. It is based on the recognition of human voice. But people's pronunciation varies from person to person, from urgent to slow, from high to low. The ambiguity of pronunciation is so great that it is difficult for robots to recognize it.

In recent years, the development of intelligent robots in China has shown a rapid growth momentum, among which the representative is the intelligent customer service robot series launched by the small I robot. Using the self-developed robot platform, Xiaoyi Robot has launched high-quality and efficient intelligent customer service robot solutions for 200 large enterprises and governments, including Dr. Fetion Haibao, Jiangsu Mobile, Tianjin Municipal Government, Lenovo Group and Bank of Communications, and has achieved positive response in society. Among them, the problem solving rate of Dr. Fetion Haibao is over 98%. When chatting with Haibao, the user feels like a real person. Jiangsu Mobile's customer service robots have more than 200,000 daily visits, and the robots handle 75% of the problems, with an accuracy rate of over 90%. Henan Mobile's SMS robot processes more than/kloc-0.0 million messages every day. Humorous answers have been widely spread in Weibo and other online media.

It is a difficult problem for human science and technology to make robots understand people. Because this voice technology has a wide range of applications, people's voices can be automatically input into computers, automatically translated into languages of many countries, and exchanged information with people of different countries and nationalities. In the field of automatic control, you can use voice control box to talk to robots, which can reduce labor intensity and improve production efficiency. Researchers at home and abroad have made some achievements in this field. 1223, an American company developed a voice-activated typewriter, which can understand 40,000 isolated words. This is considered to be the arrival of the era of practical speech recognition system. 122 1 year, Chinese researchers established a speech recognition system, which can recognize 4333 Chinese sentences composed of isolated syllables and has good knowledge learning ability.

Humans have initially mastered the technology of making robots understand human language, endowing humans with highly intelligent hearing ability, and making humans adapt to cold, emotionless machines. This technology will develop rapidly and further realize scientists' higher idea of studying speech recognition technology!