Python is widely used in machine learning research because
- It is high-level and easy to use, which allows for rapid prototyping. Researcher time is far more valuable than CPU time. (Anyways, the relatively slow performance of Python is not so much an issue because the bulk of the computation is implemented in highly tuned linear algebra libraries written in lower-level languages.)
- It is a reasonably good general purpose programming language which is already known or easily learnable by many/most programmers.
- As more people (and companies) turn to Python, there is a snowball effect.
Check out the below blogs for more understanding about this: Why Specifically Use Python for Machine Learning?