Nowadays, artificial intelligence, deep learning, then computer vision is not hot, so engaged in the industry of programmers really mastered the so-called deep learning? For this definition is really very fuzzy, but for the research of enterprise and local research institutes, boundary definition is very clear, full of deep learning theory (I think most people in industry to reach such a solid theoretical basis), so in-depth study for school students, what more important is the basic theory study, and making flexible use, run the program, the participation is more important? But if it all, the general students hard to do, and what is enterprise to the requirement of our students, we still no know,
After a period of deep learning, I found that the deep learning algorithm, the difficulty is not the ideas of the paper, but the processing of data set, the label that definition, skillfully use tensor, the steps and training network, and there is no need to consider, each layer of the BP, we are in the use of tools to build more, so obviously significantly reduces the threshold of deep learning, that reduces the threshold, then its value is more and more low, but why still fire now? May be too hot, is equivalent to flow stars,
Computer vision paper emerge in endlessly, students also in crazy hair paper now, companies will also send top papers students is limited, this kind of top talent is very valuable, but more is seldom contribution away 5,
There are more and more software, clearly not used algorithm of artificial intelligence, or there is no need to use AI algorithms, however, was for the above to depend,
Big heat, cool, cool the same arrival will freeze to death a large number of people, I feel I am ah ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ don't know what do you think you? For college students should if you want to engage in the industry, also how to prepare, how to learn, how to set their own direction?