cool hit counter What are the current mature applications of deep learning?_Intefrankly

What are the current mature applications of deep learning?

Deep learning is not a new technology, it has a long history

Deep learning was first proposed by Hinton et al. in 2006, and the essence of deep learning is artificial neural network, which is the ability of machines to learn and thus have intelligence by simulating biological neural nets.

It has also been 12 years since deep learning was conceptualized, and artificial neural networks were proposed long before the concept of deep learning was introduced.

1956 John. McCarthy introduced the concept of artificial intelligence.

And before the concept of artificial intelligence was introduced, the theory of neural networks as a computational model was originally proposed by scientists Warren McCulloch and Walter Pitts in their paper in 1943.

The real sense of artificial neural network was born in 1957, when Frank Rosenblatt, a professor at Cornell University, introduced the concept of "perceptron" and defined the world's first neural network through an algorithm.

So deep learning actually has a long history, long beyond half a century, and is not new, but the practical applications are out there and in some areas are already very mature.

Face recognition is arguably the most mature application of deep learning today

Image recognition is arguably the most favored application direction in the field of deep learning or machine learning, both deep learning and other machine learning methods (such as VM i.e. vector machines) are very fond of using image recognition for research purposes, probably because things like images are more easily accessible.

It is for this reason that the ImageNet competition was created, and ImageNet invites major AI researchers to come and test the effectiveness of its algorithms by building a powerful library of images.

Because of this, a large number of researchers have been attracted to the field of image recognition, which has contributed to the boom in this field and led to the fact that today, face recognition is almost the most mature application in the field of artificial intelligence.

How far are other deep learning applications from us?

While face recognition is said to be applied very successfully (and more accurately than human eye recognition) through deep learning, other areas are not necessarily applied as well or the results have not yet surpassed humans.

1) Driverless

Driverless technology involves very many aspects, can be said to be a comprehensive artificial intelligence technology, deep learning is only one of them, the current driverless technology, to break through a lot more, the biggest obstacle is afraid that human trust in the safety of driverless may take a longer time, personal estimates need 5-10 years to really see driverless cars running on the road.

2) Speech recognition

Speech recognition is now so well developed that basically even local dialects can be recognized very accurately and are mature applications.

3) Natural language processing

The natural speech piece still has a long way to go, and there are still many unbroken areas of machine understanding of natural speech.

4) Robotics

Although more and more robots are now available, most of them are still in the doll stage and there is still a long, long way to go for true intelligence, and it may also take 10-15 years to make a big breakthrough.

Fewer and fewer publics are still sticking to originality

One of them is "WENSEI SAYS TECH"

1、Front and backend separation in practice
2、Do you believe it The Associated Press already writes with robots
3、Leetcode punch card No18Sum of four numbers
4、Classic questions for frontend job interviews super complete
5、Diagnosis of the Vitality of the Eastern Speciality Towns

    已推荐到看一看 和朋友分享想法
    最多200字,当前共 发送