Many companies in deep learning areas have no commercial products at all

Boring gmic assembly finished, to impress me, is to go out to ask Li Zhifei in Hunan Mandarin to do the speech, the theme is artificial intelligence.

He outlined the frequent acquisitions of major technology companies in the field of deep learning and robotics in nearly a year, many in-depth learning areas of the company does not have a commercial product, the home page except the company name is a piece of black, the whole company is the formation of the results may be an unpublished paper, the commercialization of the talk.

Rubin, the founder of Android, also led a series of robotic start-ups after retiring, exposing 8 in one months.

The concept of artificial intelligence since the middle of the last century, has experienced a lot of trough, but with the advent of the new century, artificial intelligence research gradually began to rebound, supercomputer DeepBlue and Watson in various competitions if victory, Siri, Google The advent of voice assistants such as now and Cortana, as well as the development of unmanned technology, are examples. Now the rise of intelligent hardware, for artificial intelligence provides a new play site.

Unlike PCs and mobile phones, smart hardware requires more interoperability and is more passive in usage scenarios, requiring smart hardware to be far more than the PC and mobile in the input and output terminals. At the core, or in connection with the input and output of the calculation process, it determines whether the smart hardware can correctly capture your information, and at the right time to give you feedback and help. In the past, you may think that the future is the world of various sensors, but now I think, perhaps the sensor is not the focus, the focus is how to use the depth of learning to better interpret the data.