Typically, your intentions should determine which programming language you should pick up as each language (without a loss of generality) provides some benefits over others for specific types of tasks.
However, since your question is specifically related to Deep Learning, I would recommend Python as it is the most trending language for Deep Learning. The reasons for this are:
Python is centered around usability and readability and is hence the programming language of choice for rapid development.
It has an extensive set of libraries and frameworks targeted at a wide variety of users. These include web development, scientific computing, machine learning, etc. Want to deploy your deep learning solution on the web? No problem! Build the core functionality in TensorFlow and use a Django server!
Some of the most trending libraries for deep learning such as TensorFlow, Caffe, Keras, Torch are available for Python.
In the recent years, there have been tremendous efforts made for allowing developers to tap the potential of GPU computation, and optimization of core libraries for better performance. Python has tremendously benefited from this. One recent example that I can provide is Intel's DAAL [[ https://software.intel.com/en-us/blogs/daal ]]. Relating this to point 3, it is the right time to learn Python.
Also, TensorFlow comes with extensive tutorials aimed at users of different levels of expertise. You can learn DL and immediately relate whatever you've learnt to code by following their examples.
Typically, your intentions should determine which programming language you should pick up as each language (without a loss of generality) provides some benefits over others for specific types of tasks.
However, since your question is specifically related to Deep Learning, I would recommend Python as it is the most trending language for Deep Learning. The reasons for this are:
Python is centered around usability and readability and is hence the programming language of choice for rapid development.
It has an extensive set of libraries and frameworks targeted at a wide variety of users. These include web development, scientific computing, machine learning, etc. Want to deploy your deep learning solution on the web? No problem! Build the core functionality in TensorFlow and use a Django server!
Some of the most trending libraries for deep learning such as TensorFlow, Caffe, Keras, Torch are available for Python.
In the recent years, there have been tremendous efforts made for allowing developers to tap the potential of GPU computation, and optimization of core libraries for better performance. Python has tremendously benefited from this. One recent example that I can provide is Intel's DAAL [[ https://software.intel.com/en-us/blogs/daal ]]. Relating this to point 3, it is the right time to learn Python.
Also, TensorFlow comes with extensive tutorials aimed at users of different levels of expertise. You can learn DL and immediately relate whatever you've learnt to code by following their examples.
I completely agree with M. Basak recommendation concerning Python as the language of choice for deep learning.
You could also look at this description of the pros/cons of Python compared to other languages. What other language you already master might also influence your choice as discuss there:
Let me add what I have done when I encountered the same problem.
I am very comfortable of using c/c++ and Java. As Suryoday Basak said, I found most of the Deep learning based libs and tools support python. So I used a video tutorial to learn basics of Python and then practiced it for 1-2 days. It took me only 3 days to get very comfortable and confident with python. I wish the story may be helpful to motivate yourself to learn python.
Depende of the problem that be want develop. For example, if you want perform signal processing, it is recomendable use a lenguaje that be easier processing these types of problem, e.g., Matlab. If want a performe in executión C could be a option, or for graphics there are a lot. First, analize the problem and choose the better option.
Dear Di Yuan, I agree with the given answers in favour of python as a solid base (not only for machine learning).
With respect to machine learning with python you can use the TensorFlow framework (https://www.tensorflow.org). Other python essentials are the scientific computing package NumPy (http://www.numpy.org), matplotlib (https://matplotlib.org) for data visualisation and the data science platform under https://www.anaconda.com (also includes machine learning). Lots of scientific programs support customised python add-ons - i.e. your own python code (e.g. IBM SPSS, and many others). Enjoy python!
I think Depending of the problem that be want process. For example, if you want perform image processing, it is recomendable use e.g., Matlab. If want a performe in executión C could be a option, or for graphics there are a lot. First, analize the problem and choose the better option. I agree with Juan José Gómez Ricardez
If you're on a mac or linux box, Python. It comes installed. You can easily install dependencies to make a project work. You can git clone open source projects and use them out of the box. There are tools that expose machine learning algorithms through an API like scikitlearn, and lower level tools like pandas for working with data sets and matplotlib for visualization. Also there's jupyter notebooks which literally teach you how to do ML. The Python ecosystem gives you everything you need to get started. Its like training wheels.