Graduate, can not find the knowledge if a start point, then combine research mentor and the subject of design knowledge to the field, tied the basics ability. Natural language processing field of my research, for example, if a start point do not know the research, and have to arrange learning program, then I'll choose:
(1) to see the contents of the previous research senior senior sister apprentice, the mentor, to identify knowledge involved.
(2) by involving knowledge, learning positioning range, namely: Python programming language; Python combat natural language processing, machine learning practical; machine learning, statistical learning, deep learning. These books will master: Python programming; NLP is about how children, how to do in Python; in natural language processing algorithm How and what are the commonly used algorithm; what are the algorithms that are What are the characteristics.
The above two points would be too good to enrich the content of Kenichi, Kenichi wait until the end of the period, semester or study, not to know a little research, but also in the process of doing research, because a solid grasp of the basics, you can use the characteristics of data to select the appropriate algorithm, later had the idea can also be due to a solid Python programming ability, making use Python to know what the library, what code does what function when writing code.
Note: Do not think the study did not find a point to learn a wide range of knowledge is a waste of time.