[gprime]

At the moment I'm researching AI. I am noticing that there a lot of terms and definitions out there.
What I would like to make (at least for myself, maybe also interesting for others   ) is a clarifying presentation and/or schematic of all (or almost all) terms used in the AI field with their description, relationships, overlap, possible hierarchy and differences.
1. I am wondering if their are any other attempts made to do this?
2. Which definitions, term description or figure would you recommend I include in my research?

Your input is much appreciated.

With regards
Gprime

PS Until now I have scanned NG's courses on cousera(Stanford), Kampakis book about data science and some popular books like AI for dummies.
 

 

[MagnusWootton]

You may not like my answer, so dont think its important or anything, its just my wierd opinion, cause I'm a bit of a troll but I'm a real developer..

I'm aware of the terms Artificial Intelligence, Machine Learning, Deep learning,  but they don't hold much importance to me to anything specific,   the actual networking machine being used specifically is more important.

I dont use boltzmann machines, hebbian networks, gradient descent, support vector machines, recurrent neural networks, hopfield nets or even feedforward neural networks. 

I prefer to use K-Nearest neighbour, which is a form of feedforward neural network because its alot easier, conceptually basic.  (Its like the BASIC of machine learning.) and already gives good results for image recognition and speech recognition. and I haven't really moved on from there for a long time, cause it already suits my purposes.

Making KNN really kickbutt depends on what framework its being used in.  It depends on what u do with it.
And I can do a very good motor development tutorial too,  for making virtual creatures animate.

Its the actual machine you use which will entail the properties strengths and weaknesses and possible optimizations of the system.

I don't know what Datascience is and what CI even stands for.

2c.

[gprime]

Ok thank you for your input.

I'm also a developer, of old anyways (Assembler, Fortran, Pascal, C, C++). I don't program code anymore for work,.(only for hobby some assembler on my old C64 but that's beside the point). In AI field I did some expert system programming and some pick&place robot software engineering  (>30 years ago ).

For the research I do, I study some courses on ML and DL to get some idea, not to use it in practice so the stuff you mention make some sense to me  . In the current course  they are talking about FP/BP in NN and next week or so we do some SVM, on a real basic level of-course)

CI stand for Computational Intelligence (also soft computing like f.i. evolutionary computing) but that is exactly what I want to clarify, all those terms. Data science is a discussion on it self, there are a lot of people who call themselves data scientists but it is a quite broad term and can mean a lot of different types of work.

Further on-topic:
The goal of my research is to get grips with terminology and be able to explain it to people with no AI knowledge who are confronted with AI. Like business owners  and managers within companies and government.
From your reply I digest that you mean that one way to look at the broad AI field is the end-use of the AI application, which I think is a valid point and should be one part/view of the schematic that I want to make. My only questions is than what types of application are we all talking  about. Maybe Magnus you can mention a few? Or clarify how you see that? Thanks in advance.

Greetz,
Gprime

[infurl]

Welcome to the forum Gprime.

Until someone figures out what intelligence actually is, we probably won't be able to define artificial intelligence either, and a definition of consciousness is even more elusive. Sentience and sapience are also terms that are often associated with intelligence.

Check out the Free Energy Principle and see what you make of that.

[HS]

The free energy principle has been criticized for being very difficult to understand, even for experts.

Ok, I feel better now. Does Karl Friston give any good talks about this? I saw he was a guest on Lex Fridman's podcast, but I haven't listened to it yet.

[infurl]

The free energy principle has been criticized for being very difficult to understand, even for experts.

Ok, I feel better now. Does Karl Friston give any good talks about this? I saw he was a guest on Lex Fridman's podcast, but I haven't listened to it yet.

Haha, in that sense it's no worse than quantum physics and very little of the technology that we depend on nowadays would exist if quantum physics was wrong. I listened to the Lex Fridman interview with Karl Friston a while back and found it to be very helpful in putting things in perspective. Dr Friston pointed out that while the Free Energy Principle is very good at explaining a huge number of observations in the life sciences, we haven't really thought of an application for it yet. Again, that is quite normal. Amazing breakthroughs in mathematics typically take 50 years or more to have much impact on civilization because it takes generations for new ideas to take hold.

[gprime]

we probably won't be able to define artificial intelligence either

I don't agree. We can sure try to 

Artificial intelligence (AI) refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. The term may also be applied to any machine that exhibits traits associated with a human mind such as learning and problem-solving.

Artificial Intelligence (AI) as the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Artificial general intelligence (AGI) is the hypothetical ability of an intelligent agent to understand or learn any intellectual task that a human being can.

The essence of intelligence is the principle of adapting to the environment while working with insufficient knowledge and resources. Accordingly, an intelligent system should rely on finite processing capacity, work in real time, open to unexpected tasks, and learn from experience. This working definition interprets “intelligence” as a form of “relative rationality” (Wang, 2008)

[MagnusWootton]

Further on-topic:
The goal of my research is to get grips with terminology and be able to explain it to people with no AI knowledge who are confronted with AI. Like business owners  and managers within companies and government.
From your reply I digest that you mean that one way to look at the broad AI field is the end-use of the AI application, which I think is a valid point and should be one part/view of the schematic that I want to make. My only questions is than what types of application are we all talking  about. Maybe Magnus you can mention a few? Or clarify how you see that? Thanks in advance.

We all have a different grand scheme what we'd like our Ai to do for us;   myself I'm not into the AGI so much rather than just performing simple tasks but in a real physical machine,  not just game A.I. in the virtual realm.

Not that I completely am negative about AGI, I just think there's some important instrumental goals along the way that I feel need to be fully implemented before continuing on for the "holy grail".

Because the deficiencies of wheel based locomotion for being inside the buildings, not just the outside on the road,  I've spent alot of time on locomotion of legged machines,  and a cool thing is,  id like to say picking something up with an arm, is very similar to pushing a body along the ground with legs.

My end product is a robot that follows simple commands to perform simple tasks,  which isn't new in itself,  its probably a common goal since the 70's, but I want to do it very efficiently, and user friendly, for 2021.   more than what we've seen till today.

[HS]

@infurl. That was quite an interesting podcast, many novel (maybe original) ideas in there. 

[Don Patrick]

The goal of my research is to get grips with terminology and be able to explain it to people with no AI knowledge who are confronted with AI. Like business owners  and managers within companies and government.

In that case you may like to check out my laypeople's glossary of AI jargon. The degree of cheek in some of the definitions is representative of the amount of fruitless discussion that they have been subject to. Here is a similar glossary that I also found well written.

[LOCKSUIT]

"Until someone figures out what intelligence actually is, we probably won't be able to define artificial intelligence either,"

what a cheese cake attitude
I got your whole AGI 15-page-book all polished and finished! I'm working on the code now. And take a look at openAI.com to see stuff you'll never forget.

To be honest, unless you are REALLY devoted to understanding AGI and building AGI, forget about my book.

[MikeB]

At the moment I'm researching AI. I am noticing that there a lot of terms and definitions out there.

These are all the same thing:
Solid-State AI, Symbolic AI, Good Old-Fashioned AI, Rule-based system, Expert System

How they're used may be different, but they still all start with literal data that is searched on, then some other literal data is output... (no algorithms or predictions, or fuzzy logic.)