NLP, or Natural Language Processing, is an important branch of artificial intelligence that enables computers to understand, process and generate language like humans. Search engines, machine translation, and voice assistants are all powered by the technology.
The content of the field of artificial intelligence is cumbersome and complicated. It is impossible to achieve such a high level of technology in just a few months, but as long as you can follow the learning methods and suggestions I mentioned below to learn step by step, it is guaranteed that you can use the shortest time. , Getting Started with NLP!
1. Learning method
I personally think that NLP learning can be divided into two types .
The first is to start with theoretical knowledge . For example, beginners in natural language processing need to understand NLP language and some basic algorithms, such as classification, integration, dimensionality reduction and other algorithms.
In addition, the amount of code in the initial stage of learning is essential. Python is a good entry-level language, relatively easy to use, and very friendly to beginners. For later exercises, you can find related projects on Jupyter and github.
Generally speaking, this kind of learning method will take a long time, but relatively, it can also lay a solid foundation for future learning.
The second method is to start with practical cases . People who have really entered this industry should have the same feeling, that is, practical operation is much more important than theoretical knowledge in textbooks.
As a big branch of artificial intelligence, NLP covers a lot of content, professional courses include machine learning, numerical analysis, linear algebra
Advanced numbers, matrix theory, etc. The main training is code engineering ability and algorithm analysis ability to solve problems.
Such a huge knowledge system requires a lot of time to study systematically, so for this field, I do not recommend self-study, especially the way of learning by reading books alone.
Personally, I think you can download a NLTK (Natural Language Toolkit) first , and then try to build some software, such as: run some different POS taggers
And try to describe the differences, see which domain they fit better or what similar mistakes they make; build a sentiment analysis system and figure out how to make it better, what features you should use to help it do better ; Write a system that can detect spam comments on websites, etc.
2. Learning suggestions
Regarding NLP learning, here are some specific suggestions of mine :
- You must first have a solid algorithm foundation. This step should be mastered before starting NLP learning. You can learn any language, but python is the most recommended here.
- Learn regular expressions. This step is difficult to get started, but once you start learning, it is easy to learn and still follows a certain logic.
- NLP includes many concepts, such as pos tags, symbolization, tokenization, etc. And the key here is tokenization. Depending on the content of the project being built, the content of the mark is also different. In this regard, python has a good API to complete these projects.
- ML learning. The basic knowledge of ML can be selected from coursera courses. For the language of python, nltk is mostly used.
- Evaluation metrics for learning. This part is important but often overlooked. ROC curves, F-scores, precision, bias and variance, etc. all apply here.
- deep learning. After all steps, if your classifier still has high bias, or needs more complex content, this needs the help of deep learning. Deep learning is very useful for specific tasks in highly nonlinear feature spaces.
3. Learning resources
In addition to finding the right learning method, you also need to combine the teaching materials that suit you. There are so many resources about NLP learning on the Internet, and it often feels dazzling when choosing.
The following are some screenshots, get learning materials and pay attention to the official account: Gupao AI, reply: 168.
Table of contents
1. AI Free Video Courses and Projects
2. Artificial intelligence must-read books
3. Collection of Papers on Artificial Intelligence
4. Machine Learning + Computer Vision Basic Algorithm Tutorial
To learn artificial intelligence well, you need to read more books, do more hands-on work, and practice more. If you want to improve your level, you must learn to calm down and learn systematically slowly, so that you can gain something in the end.