Recently the company to do machine learning and artificial intelligence

Namely through the study of large data forecast data

Simple learning ml.net suddenly feel good,,,, feeling distant and artificial intelligence suddenly before I can do it,,,,

Is my illusion?

I was afraid that a problem,,, I very much doubt I understand not enough in-depth or ml.net is so simple,,,,,,,,,,,


Who has been using a discussion

CodePudding user response:

Well, you mean "python please pick up the knife, are you ready to accept fate?"

CodePudding user response:

This is the important learning and prior learning, playing with the python itself usually does not explain why his program to meet the target, can only say that can happen to like this, finally from the probability and statistics, at least he was "pretty accurate"

So afraid of a problem? Does not exist, because so far as to the subject is a priori, unable to say why, so there is no fear of anything out of the question, because there is no "problems" and "no problem" clear boundaries (yes, you play this, of course, understand that threshold is what, why, of course, this threshold, you ask the python, python also don't know, he will answer you this threshold, behave better from the probability and statistics, so you can't to say the obvious boundary problem, this problem in this does not exist at all)

CodePudding user response:

Common scenario can processing with ml.net, really very good,
Machine learning has developed over the years, we can use to the scene, have been categorized to build good solution framework, the more contact later, the more likely they were to say simple, just like before driving is automatic, all kinds of shift operation, the novice hands busy feet; Now most car is automatic, out of the box, many common operations are built in before they go out, don't need later users care about,
Most of the scene has a model, only need to provide data for training is ok; Even said that many models have been training well, directly in line with it, such as 12306 verification code identification model of speech recognition model, model of image classification, etc.,
If you to design a artificial neural model, that is a lot of things to learn, some of the concepts I watched many times also don't understand, also some of the high number of things, and that kind of guesswork,

CodePudding user response:

Learn to 1 + 1=2, also need not to study why 1 + 1=2

CodePudding user response:

Tried to use ml call tensorflow trained model.