News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
What will happen to AI and humans in the future?
What will happen to AI and humans in the future?
At present, the development of robots is getting faster and faster, and there are different voices that come with them. In the future, robots dominate humans? Or do humans dominate the robot? I think everyone understands that human beings must dominate robots. Why? Because of the words.
Is the future a robot or a human being?
If you want to enter the artificial intelligence category, you should first learn Python.
Although the artificial intelligence category supports many other languages, Python is the most widely used and simplest one. But why choose Python - after all, is Python so slow? Since most libraries use the symbolic language approach rather than the imperative language approach. Explain that is to say: instead of executing your instructions one by one, create a computing graph based on all the instructions you give. This diagram is internally optimized and compiled into executable C++ code. This way you can apply the best of both worlds: the speed of development that Python brings and the speed of execution that C++ brings.
I often hear people talking about deep learning and artificial intelligence - where should I start? TensorFlow is the most popular today? I heard that Caffe is very common, but will it be too difficult? At BEEVA Labs, we often need to deal with many different deep learning libraries, so I hope to share our findings and feelings to those who have just entered the field of deep learning and artificial intelligence.
First, let's talk about a few commonly used deep learning languages.
TensorFlow
TensorFlow is an open source software library that uses data flow graphs to stop numerical calculations. TensorFlow is Google Brain's second-generation machine learning system, which was once open source. TensorFlow can be applied in many central areas, such as speech recognition, natural speech understanding, computer vision, advertising, and more. TensorFlow is a very sensitive framework that can run on a single or multiple CPUs and GPUs on a PC or server, so that it moves on the device.
TensorFlow supports both Python and C++, as well as computational dispersal on the CPU and GPU to support gRPC stop-degree extensions.
Caffe
Caffe is not just one of the oldest frames, but an old card. At first it was not a general framework, but focused on computer vision, but it was very versatile. In our lab experiments, the CaffeNet architecture exercise time was five times less in Caffe than in Keras (using Theano backend). The drawback of Caffe is that it is not sensitive enough. If you want to give it a new change, you need to use C++ and CUDA programming, but you can also use the Python or Matlab interface to stop small changes.
Caffe's documentation is very poor. You need to spend a lot of time checking the code to understand it (what is the use of Xavier initialization? What is Glorot?). One of Caffe's biggest drawbacks is its device. It needs to handle a large number of dependencies...
MXNet
Mxnet is one of the frameworks that support most programming languages, including Python, R, C++, Julia, and more. But I think developers who use R language will especially prefer mxnet, because Python has so far dominated the deep learning language with an unquestionable situation. I have some doubts about the expansion of multi-GPUs and I am very interested in understanding the details of such an experiment, but I am still suspicious of mxnet.
So why is python suitable for artificial intelligence?
Google's TensorFlow basically has all the code in C++ and Python, and other languages are just a few thousand lines. If you talk about the speed of the operation, use C++, if you talk about development efficiency, use Python, who will use artificial intelligence such as Java, which is not high and low? Python is a script language, but because it is easy to learn, it quickly becomes a tool for scientists (MATLAB can also do scientific calculations, but software requires money and is expensive), thus accumulating a large number of tool libraries and architectures, and artificial intelligence touches a large number of Data calculation, using Python is very natural, simple and efficient. Most of today's deep learning frameworks support Python. Who is not using Python? Python has a lot of excellent deep learning libraries available, and life is too short to use Python.
How can I use the machine learning model developed by Python to run quickly and cost-effectively, and is deeply compatible with TensorFlow? You can learn about Google's AIY Projects project. It is also promoting the artificial intelligence of developers and DIY communities through programs such as the “AIY Project”, which themselves represent artificial intelligence. Google's goal is to make AI complete the real civilian, let artificial intelligence everywhere, everyone can learn.
Is the future a robot or a human being?
If you want to enter the artificial intelligence category, you should first learn Python.
Although the artificial intelligence category supports many other languages, Python is the most widely used and simplest one. But why choose Python - after all, is Python so slow? Since most libraries use the symbolic language approach rather than the imperative language approach. Explain that is to say: instead of executing your instructions one by one, create a computing graph based on all the instructions you give. This diagram is internally optimized and compiled into executable C++ code. This way you can apply the best of both worlds: the speed of development that Python brings and the speed of execution that C++ brings.
I often hear people talking about deep learning and artificial intelligence - where should I start? TensorFlow is the most popular today? I heard that Caffe is very common, but will it be too difficult? At BEEVA Labs, we often need to deal with many different deep learning libraries, so I hope to share our findings and feelings to those who have just entered the field of deep learning and artificial intelligence.
First, let's talk about a few commonly used deep learning languages.
TensorFlow
TensorFlow is an open source software library that uses data flow graphs to stop numerical calculations. TensorFlow is Google Brain's second-generation machine learning system, which was once open source. TensorFlow can be applied in many central areas, such as speech recognition, natural speech understanding, computer vision, advertising, and more. TensorFlow is a very sensitive framework that can run on a single or multiple CPUs and GPUs on a PC or server, so that it moves on the device.
TensorFlow supports both Python and C++, as well as computational dispersal on the CPU and GPU to support gRPC stop-degree extensions.
Caffe
Caffe is not just one of the oldest frames, but an old card. At first it was not a general framework, but focused on computer vision, but it was very versatile. In our lab experiments, the CaffeNet architecture exercise time was five times less in Caffe than in Keras (using Theano backend). The drawback of Caffe is that it is not sensitive enough. If you want to give it a new change, you need to use C++ and CUDA programming, but you can also use the Python or Matlab interface to stop small changes.
Caffe's documentation is very poor. You need to spend a lot of time checking the code to understand it (what is the use of Xavier initialization? What is Glorot?). One of Caffe's biggest drawbacks is its device. It needs to handle a large number of dependencies...
MXNet
Mxnet is one of the frameworks that support most programming languages, including Python, R, C++, Julia, and more. But I think developers who use R language will especially prefer mxnet, because Python has so far dominated the deep learning language with an unquestionable situation. I have some doubts about the expansion of multi-GPUs and I am very interested in understanding the details of such an experiment, but I am still suspicious of mxnet.
So why is python suitable for artificial intelligence?
Google's TensorFlow basically has all the code in C++ and Python, and other languages are just a few thousand lines. If you talk about the speed of the operation, use C++, if you talk about development efficiency, use Python, who will use artificial intelligence such as Java, which is not high and low? Python is a script language, but because it is easy to learn, it quickly becomes a tool for scientists (MATLAB can also do scientific calculations, but software requires money and is expensive), thus accumulating a large number of tool libraries and architectures, and artificial intelligence touches a large number of Data calculation, using Python is very natural, simple and efficient. Most of today's deep learning frameworks support Python. Who is not using Python? Python has a lot of excellent deep learning libraries available, and life is too short to use Python.
How can I use the machine learning model developed by Python to run quickly and cost-effectively, and is deeply compatible with TensorFlow? You can learn about Google's AIY Projects project. It is also promoting the artificial intelligence of developers and DIY communities through programs such as the “AIY Project”, which themselves represent artificial intelligence. Google's goal is to make AI complete the real civilian, let artificial intelligence everywhere, everyone can learn.
PREVIOUS：Face recognition self-service machine
NEXT：Passport OCR