Artificial use of intelligence

The Artificial Use Of The Term "Intelligence"

There are a great many mistakes made in recent articles when it comes to terminology in the A.I. sector. Let's examine this, and see what we can find out!

A great many articles are written every day about everything to do with artificial intelligence, machine learning, deep learning, etc.
Sometimes these terms are used with great care, and applied in exactly the right context, other times not so much.
If you are very familiar with this field, eiter through practice or study, you may already know exactly what I am talking about, and might even be able to correct misuse of terms on your own, but for those of you who want to get a finar grain understanding of what is going on, keep reading.

SIGN UP FOR OUR NEWSLETTER

The newsletter will keep you informed of interesting projects in the fast-growing field of A.I. and machine learning.
Of course, we promise we will not spam you, and no shameless self-promotion.

PRO TIP: Highlight any text to share a quote easily!

The Infamous Subset Example

Maybe you have seen this image on the left, or some variation on it. It is shared quite frequently by people on social media who have not much else to share but infographics, and fluff pieces written by journalists.
There are just a couple of things wrong here, in my opinion.

Here's the deal: Starting from the top I think it is crazy to say that artificial intelligence must mean computers reasoning as humans.
First of all reasoning as any other animal would be more than acceptable from my point of view, and secondly reasoning as something completely separate from any biological entity can still be considered intelligence.

Let's move on, and look at the machine learning layer, which is probably the best term to use for the technology must of us read about on a daily basis.
I think the infographic boils it down to its absolute simplistic core, which it is meant to do of course, and is inherently not wrong.
The only thing that I am not 100% convinced of is its position within the A.I. circle.Think about it, is machine learning truly a subset of (artificial) intelligence?
While it is true that the algorithms currently used rely mostly on learning from input data, very little to no reasoning at all is going on once the model is trained, or even during training.
In fact, machine learning algorithms are basically just huge composite functions, and with some additional tricks of advanced models aside, perform simple addition of the inputs and previous layers, multiplying the weights.

It is these weights on which the training is performed, so in very simple terms, you run your training data through the untrained network, and calculate the error of your outputs based on your expected outputs, then you adjust your weights based on your error value.

Why So Sensationalist?

So, to shield myself from clever people who might want to take the time to go through my older articles to find myself doing to exact same thing as I am preaching against in this piece: Yes, I have absolutely been guilty of peppering in the wrong terms myself.
I use artificial intelligence, machine learning, and other words almost interchangeably, and I really have no excuses to offer here, but let me try nonetheless.

I think one of the main reasons myself, and most other blogs writing on the subject do this not just because they are uneducated on the correct use of terms in the field, as terms are easily learned.
There are however certain practices around search engine optimizations to consider, which is kind of funny actually, because it seems that modern machine learning may be forcing us to write incorrectly on the topic.
For those who understand this: Word2Vec might be fun, but it isn't very accurate (an article on that coming out very soon).

What I don't understand is why we do not just use the term "machine learning" for every case were we are talking about machine learning. It is not rocket science, but I guess that takes away some of the mystique that we feel when we talk about artificial intelligence.
I see a lot of articles and discussions being had that make me wonder if most of the world is actually taking the films they have seen as a serious spring-board into this discussion.

In fact, I know this is true, because many times when I talk or write about safety considerations in (true) artificial intelligence, people will bring up Asimov's Laws of Robotics, and that should tell you enough.

But, What About My Headlines?

Of course, as I mentioned earlier, the writers out there still want to be able to benefit from the sensationalist qualities that some terms can bring over others, but I will argue that by correctly using the right terms should not limit anyone to write headlines that can still grab attention.
The fact is, one can still use artificial intelligence in a headline, you will just have to write in a different scope within the corresponding article.

By educating yourself on the topic you are writing about, you will in fact have way more to offer in this niche.

Personally I am going to try from now on to stick to my new ideas on where the differences between terms, and the dividing lines are, and hopefully stay consistent, yet morph with any new incoming information.

Follow me on social media to get updates and discuss A.I., and I have recently created a slack team to see if people are interested to forming a little A.I. community,
Twitter  : @ApeMachineGames
Slack    : join the team
facebook : person / facebook group

Ai

Hacking in the World of Artificial Intelligence

A look at hacking in the world of artificial intelligence, and how to trick machine learning algorithms.

Humanai

Human Compatible A.I.: A Myth

While the term "human compatible artificial intelligence" is used a lot by researchers and thinkers in the field, I believe this whole concept to be an absolute myth.

Optmize chatbot

5 Simple Ways To Optimize Your Chatbot (HowTo)

If you have ever messed around more with chatbot services like api.ai or recast.ai, and got beyond the "order a pizza" example, you might be wondering: Why does my chatbot start messing up when I train it more?

We all know the scenario, starting with a simple bot, and adding the small-talk intent to it that all of them provide out of the box. Things are looking great so far!
Adding more intents and features to our bot is working out too, and we are soon dreaming of building a general solution to all of our customer service related problems.

What's more, many companies are sprouting up targetting businessnes big and small, promising them solutions for customer service, employee onboarding, and other tasks that sitll require some human finesse.

Sadly, things are not that simple...