Here we finally are: Artificial Intelligence. 🚀
Let’s go straight to what matters to us; in this chapter, we will launch firstly into a knowledge more advanced about AI, and we will delve into some tools secondly, before discovering in detail the GP3 tool thirdly, and the Coh-Metrix tool fourthly.
In that way, we will be training to not fall in traps despites ourselves and to enforce our knowledge concerning the word weight.
- AI: historical, target, purpose, capabilities
Difficult to know how to start or where to begin. So, let’s get to the heart of the matter with some watching and reading to make.
You can look this article over to have an insight of AI and its historic: https://www.ibm.com/cloud/learn/what-is-artificial-intelligence
The passage with the Key Dates and names is very clear and summarises well the chronology of.
And also this video, 9 min though, but very well explained and interesting: https://www.youtube.com/watch?v=kWmX3pd1f10
So, as well as the article and the video, we talk about “intelligence”: and the question asked during the introduction comes out from our mind again. Is a machine able to think ? Because if the thinking would precede the language as mentioned in the introduction, what is it concerning AI: could they be intelligent and have a kind of language, without thinking beforehand?
Let’s cast a glance over this article, that tries to answer to this crucial question:
In the previous chapter, we talked about the manner and way to use AI: counter Fake News generated by AI, with AI. Use the same weapon, but with a purpose diametrically towards another end.
In the same vein, because morality cannot be extracted from these questions, a concept is born: the one of Friendly artificial intelligence; that means get position in the use of AI in a friendly way. The page wikipedia corresponding is quite relevant, speaking of which: https://en.wikipedia.org/wiki/Friendly_artificial_intelligence
This concept is in straight relation with the morality we put behind our way, because it supposes to bring the use of AI into line with our intention.
In this idea, there is a quite interesting sequence, still in 2001 directed by Stanley Kubrick, that depicts how much the AI robot, Hal9000, has capabilities that come under more of interpretation than just completion of its tasks. It’s a sequence during which the two main characters decide to keep themselves to themselves, standing apart from Hal9000 because of suspicions concerning it, but Hal9000 succeed to figure out their conversation, just by reading upon their lips. And more than just depicting the astonishing potentiality of this AI, it is also a sequence that illustrates the tiny boundary between the good use and the bad use of AI; and the fears that one day, AI could overtake human beings, in a morbid desire for power.
Please, watch the sequence in question in the PowerPoint.
And some institutes like this one https://allenai.org base their philosophy and common thread on this philosophy: in the service of the common good.
AI has many capabilities; and you have already noticed that because during the two previous subsections, you used some tools generated by AI to help you counter Fake News.
But, what we didn’t showcase yet, are the capabilities of AI regarding the self-improvement of writers, journalists and so on.
And concerning you, you have as a project to become journalists, so to expose yourself in a field of which the main purpose is to share information. Even if manners to transmit this last can be as varying as there are people to share it, and even if you have numerous capabilities to do that very properly, you’ll never be completely protected against pitfalls and mistakes. And it is exactly in this direction that the use of AI in a friendly way suddenly becomes meaningful ; because it allows you the possibility to improve yourself, to improve your skills in regard to your writing, the choice of words, sentence structure, and so on.
And this is exactly the point we are going to figure out on the three next subsections ; in analyzing some specific tools to corroborate this idea according to which AI could help you in your projects and professional intentions.
So, let’s pursue!
- Different useful tools
Don’t you think it is a little strange to very regularly attest before entering in some websites that we are not robots, and to affirm that to a robot itself?
What is quite interesting with AI, it’s that it’s always on the way to open numerous doors to think about. And this idea of ‘mise en abyme’ comes back, again and again, continuously when we talk about AI, because the purpose of AI is exactly to simulate human intelligence.
From that point of view: imagine a machine, which is trying to pretend it has human intelligence, knowing that we are also human; this amounts to trying to speak with robots that we would like to make them almost as human as us, but at the same time a robot that asks us if we aren’t either robots ourselves…
That’s why, kind of “mise en abyme”, or “metatheatre” but adapted to a broad way of thinking, and not only suitable for the art field.
Who, among all of you, has never used Google Translation? Or another tool for translating sentences from your own language to another one?
Who, among all of you, has never used the autocorrect, included on your smartphone, or in Google Doc ; including also the capability to guess the end of a pending sentence?
Whether you know it or not, you used Artificial Intelligence all this time.
Whether you know it or not, AI surrounds your activities much more than you can think.
It is also the case with some GPS, like Waze, that uses information in real time : it is able to inform us about the traffic, potential accidents, traffic signs ; but also the other users on the road. And because the funniness, with AI, is never far away, you can also choose the voice that dictates the way, and among these voices, there is Homer Simpson’s voice, or still, your own voice that you can register.
So, there are some tools that we use in our daily life, without even noticing ; but there are also some tools we can use intentionally, for a precise purpose.
It is the case for some tools whose intention is to improve the writings of people, with a system that is enabled to recognize mistakes, to analyse the form and to help reworking and reformulating sentences to make them closer to our intention or more professional, more consensual… Or whatever else.
Some of them are: Writing Pal, LIWC, AI Writer…
Most of the time, they aren’t free, but there is a possibility to test them or use some demo.
You can try, or you can simply read about them, familiarize yourself with their intention and capabilities, to develop basics for the next, with tools a little more intricate.
Writing Pal: http://www.adaptiveliteracy.com/writing-pal
LIWC: https://liwc.wpengine.com
AI Writer:
You can also find this kind of tools, very convenient: https://www.semanticscholar.org :
It is a free AI powered research tool for scientific literature ; drawing a list of numerous scientific articles.
Or, still, this test Eai Language models: https://6b.eleuther.ai/?fbclid=IwAR3L2qI-ZeJpiT_q2KU9DZNKH6x60r1j-lf8YyzmT4XewUHLe2F6L63tY3k
You can ask some questions, and notice the answer generated by.
You can also cast a glance over this interesting Lithuanian tool:
https://6b.eleuther.ai/?fbclid=IwAR3L2qI-ZeJpiT_q2KU9DZNKH6x60r1j-lf8YyzmT4XewUHLe2F6L63tY3k, and ask some questions and see what it’s coming out from. Personally, I tried two existential questions and very broad ones, just to drive the tool into a corner ; “What is love?”, and “What is life”, and also to probe its boundaries regarding what it comes under philosophical questions, thoughts, thinking ability and so on.
Here is the answer to “What is Life?” – more relevant than the one about love – that you can take note of: “What makes an intelligent lifeform tick? The beginning of an epic journey across the universe. This is the story of a teenage girl that will venture through countless dimensions, searching for answers on what makes us tick, on the big questions, on the things we cannot understand. This is the tale of how a teenage girl and her trusty dog combine all the most powerful tools at their disposal, to get to the bottom of it all.
And that is the story of the Gem Universe. The only trilogy of its kind, and it all starts here!
Including real life experiences, with people and objects from my own life.
It might not be a novel, it might not be like anything you have read, or seen. But it’s what you are thinking. It’s your mind working, and I want to help you get to where you need to go.
And if you are thinking of getting a puppy, we should get a dog together!
Thanks!
Sara
PS: Puffin Books, you will be seeing this novel, and I’m glad.
The early version was a rough draft, and if it works out as a polished product, I’ll do a print-I’ll do a print-on-demand version to make it more available to more people!”
You can also cast a glance over this computational intelligence, “WolframAlpha”: https://www.wolframalpha.com. This AI can answer questions, and is abundant of knowledge, statistics, details and even, curiosities if you’re in the mood for that. You can enjoy discovering secret paths contained in every section. And compared to your search engine – that brings you toward either fake news or true news but clearly biased -, this website is completely neutral.
Some demos, also, drawn up in a list: https://allenai.org/demos?o=11
Sometimes, also, the usage of AI and of the statistics help to make decisions in art, as this article shared on Facebook shows: https://www.facebook.com/10513336322/posts/10160359004561323/?sfnsn=mo
None of the tools is flawless: there is always a margin of error. Something that proves that machines and AI aren’t like humans? Well, interesting question, but we must define what we hear behind the “error”, because from this point of view, human beings aren’t the exception.
Do robots could become “perfect”, more perfect than humans – with the definition of perfection in relation with the non-error?
It was a first approach, a first sight of AI with some interesting and entertaining tools ; but still not very able to provide you a very significant help, concerning the intention to enforce your knowledge.
There are more and more tools, as you can imagine, but next, we are going to look deeper into two tools, much more relevant than the ones before, because they are much more sophisticated: GPT-3 and Coh-Metrix. That you have already approached the first lines during your homework!
- GPT-3
To contextualise a little more, note that many AI research laboratories exist. One of them, American, is OpenAI and it is in competition with another, British, DeepMind.
The purpose of OpenAI is also to promote and develop an Friendly AI – the concept mentioned just before – for the benefits of humanity as a whole.
Here OpenAI website : https://openai.com
You can also cast a look at this subsection for the website : https://openai.com/blog/dall-e/. It is one of their projects with a neural network “DALL-E”, a GPT-3 parameter that enables it to create images from text-description. It’s quite funny and useful to take note of that. There are numerous examples and descriptions well provided.
OpenAI creates some Generative Models: GPT, GPT-2 and GPT-3.
It’s GPT-3 that is going to interest us, from which you read about during your homework.
You can, from now on, switch from the theoretical PowerPoint to the practical PowerPoint.
Because of this homework accomplished, you are now familiarized a little bit more with GPT-2-3
and Coh-Metric.
Now, the purpose of the hands-on concerns the discovery, by the application, of GPT-2 or GPT-3.
For helping you, you can watch some explanatories videos about GPT-3, like this one: https://www.youtube.com/watch?v=8psgEDhT1MM.
It is difficult to find some demo of GPT-3 in free access, but this one, which concerns GPT-2, could give you a first sight of what these Generative Models can imply: https://transformer.huggingface.co/doc/distil-gpt2
Spend the necessary time to attempt to test its boundaries.
Start, explore with the PowerPoint for the hands-on, and let’s move on, pursuing the next with Coh-Metrix.
- Coh-Metrix
Basically, Coh-Metrix is a computational tool that produces indices of the linguistic and discourse representation of a text […] analyzing texts on many different features”.
Do you remember this article, mentioned before, about Democracy vs Dictatorship? Well, it was just a glance, but now, we can notice that all the analyses about the language throughout the article – formal-informal, threatening and so on – have been noticed also because of Coh-Metrix.
There is the link, as a reminder: find it again.
After studying the article, we can note that it displays how it is possible through the language, and thanks to the use of Coh-Metrix, to interpret the stance of Nations – informal vs informal. One of the conclusions in regards to that, is that speaking with less formal language betrays speakers that feel themselves among the people ; whereas speaking more formal betrays speakers that feel themselves out from people, in higher status or betraying a hierarchy in the system. This complex language and syntax is also concerning countries that are “trying to project or convey their authoritativeness, stature, or accomplish greater legitimacy in the international system”.
The article also portrays, thanks to Coh-Metrix, the correlation between threatening language and anger level (with North Korean example) : interestingly and almost paradoxically, the anger level is more important when threats are just pronounced, without being accomplished, than threats executed : anger is more conveyed when threats are “held in reserve”.
Other important terms are also emphasized sometimes, like “Deep cohesion” and “referential cohesion”, bring into opposition each other ; the deep cohesion being an information broadly and thematically related, whereas the referential cohesion tends to be more memorable, useful, quotable.
What is interesting about the use of Coh-Metrix, is that it combines several aspects that we have studied since the beginning together.
Do you remember the Mental Representation and the 3I – Intention/Interpretation/Influence? Well, one of Coh-Metrix’s intentions settles on the subject of the coherence in the mind of the reader; that means that “when the level of cohesion in the text is insufficient for the reader or when the reader does not (or cannot) generate sufficient inferences to make connections between ideas, then the reader’s understanding will be less coherent”.
In other words, the cohesion inside the text is essential to enable the reader to make sense of his or her understanding, and make consistent connections. It is an intrinsic relation between text and Mental Representation, depending on one another.
The use of Coh-Metrix can be very relevant for journalists, because “processes texts for numerous indices of cohesion, language, and readability, which together allow the tool to estimate a wide range of textual features reflecting cohesion relations, and world knowledge, together with language and discourse characteristics”.
In that way, it enables future journalists as you to be helped in choosing the most appropriate words and tone as possible, according to what you want to say, but also according to who you want to be : in line with the previous analysis, concerning Democracy vs DictatorShip, did journalists want to indicate a separation, an hierarchy between them and their readers, or did they want to be involved with them, as the same whole ?
There it is the official website of the tool: http://cohmetrix.com; you can test the tool at your convenience.
Follow the instructions in the PowerPoint for the hands-on.
You have the possibility, not only to test the tool, but also to read significant documentation (more than the one read during your homework), provided with, to learn more about what it is. Read and learn are for now the priority ; because you will train just in the next chapter.
Moreover, a book’s pdf about Coh-Metrix, quoted two times in this subsection, will be also provided and is very exhaustive in the matter of; so you can look it up to deal with Coh-Metrix in depth.
Because this chapter, concerning AI and some of the tools that exist, comes to an end right now.
And a new chapter is going to open, a chapter destined for the hands-on, after digesting all this theory since the beginning of.
So, let’s move on!