This is a follow up to my previous post, where I talked about books and videos I recommend for people who want to increase their general knowledge of NLU.
This post is going to be different. I feel it is very hard to make recommendations - because I already know many people who read this website are what I call technical non-programmers.
There are three reasons why I am not inclined to make any recommendations on this topic.
- I may be suffering from a bit of an imposter syndrome on this. I know the basics, but I am not sure if I know enough to make broad recommendations
- the fields of ML/AI are growing too rapidly for books and videos to stay current for any reasonable length of time
- Donald Knuth hasn't written a book on Machine learning yet. When he does, you know that it will be more or less the final word on the topic 🙂
The four types of people who read this article
In addition, not everyone who is interested in this article is actually at the same level when it comes to skill set and coding ability.
The Technical Non-programmer
This is a person who is technical, and not a programmer. If you are only interested in business outcomes which can be generated by ML/AI, this describes you.
You can use the NLTK library (as an example) and perform simple NLU tasks like parsing a sentence and finding named entities.
The Research Programmer
Did you know that one of the most powerful NLP tools, Stanford's CoreNLP, is considered as "research exhaust" by the very folks who wrote the code? Well, I suppose it is so hard to understand the code architecture, it kinda figures 🙂
No definite ones—a lot of this comes down to happenstance. CoreNLP is primarily “research exhaust”, even if we do quite a bit of maintenence
— Stanford NLP Group (@stanfordnlp) December 5, 2016
On a more serious note, there are some folks who literally take all the core research coming out of the group and turn it into software tools which can be used by others. Usually, they have a background in computer science (which isn't true of programmers in general) which helps them understand the algorithms and mathematical notations and such.
In fact, quite often, the folks who write and maintain such code are actually graduate students or soon-to-be graduate students in Computer Science.
This would be the person who actually produces the research which is then translated into software like CoreNLP.
Someone like Donald Knuth would fall into this category, although I doubt if he needs much help learning Dialogflow 🙂
Now, depending on which category you fall into, the recommendation will vary.
As a result, I don't have specific recommendations.
Do you have a SPECIFIC question?
On the other hand, do you have a specific question about ML or AI?
You can fill out the form below, and I will answer your questions in the form of articles if and when I get some time, or send you a reply explaining why I cannot answer the question.