Skeptic News: The Other NLP

NZ Skeptics Newsletter

 

The Other NLP

Recently I've been playing with some deep learning software - OpenAI's GPT-2 and GPT-3, and EleutherAI's GPT-J-6G. These are NLP algorithms. No, not that discredited garbage Neuro-Linguistic Programming - in this case NLP stands for Natural Language Processing.

The basic idea of these recent efforts in deep learning is to take a piece of software that has been written to guess the next word in a sequence, and train it on a huge corpus of data. It turns out that the internet is a great source of natural language, and a lot of it is very easy to scrape and feed into one of these algorithms. So these pieces of code are trained on lots and lots of internet text.

This training is very processor intensive, needing thousands of hours of time on modern PCs using specialised AI chips on expensive graphics cards. However, once the algorithm has been trained, the dataset that has been created is just a few hundred megabytes in size, and can be quickly loaded into memory - the training only needs to be done once. At this point the software can be used to predict the next word in a sequence, and can keep doing this - creating whole sentences and paragraphs that actually make grammatical and logical sense. We will see below what these general NLP algorithms can do.

The same software can also be fine tuned by giving it a smaller set of data. Using its ability to put together coherent sentences, the software can emulate the data set it’s been fine tuned on. So far I’ve been working on a couple of fun skeptical projects with this, although I have more ideas.

Anyway, it’s probably easiest if I just give you some brief intros and show you the kinds of results I’ve been getting. Enjoy!

Mark Honeychurch

Hate Speech Submission


A couple of weeks ago I talked about the NZ Skeptics putting together a submission for the Ministry of Justice in response to their new Hate Speech proposals. You can read our submission on our website, but I thought it would be fun to use EleutherAI's free online GPT-J-6G deep learning model to write a submission for us. I gave the software the first few paragraphs of our real submission, and then clicked the button to guess the next hundred or so words. I then fed the result back into the algorithm so that it could create the next block of text, and so on.

Eventually the text started getting repetitive, and veering off track, but for a while what I was being given by this trained algorithm made sense and looked pretty believable. I didn't agree with its arguments, but at least it seemed to be creating text that reads like a coherent argument. Please be aware that the pieces of legislation it mentions probably don't exist - this software is just trying to emulate the kinds of language it's been trained on, writing in blogs and on forums from all across the internet.

In the text below, everything in bold comes from our actual proposal. The ensuing text that is not in bold, starting at "We also acknowledge that there is a public interest", is written by the deep learning algorithm. Have a read and see if, like me, you're more than a little bit impressed with how well it did:

 

NZ Skeptics submission on the Ministry of Justice Proposals against incitement of hatred and discrimination

 

The NZ Skeptics


We are a New Zealand/Aoteaoroa charity dedicated to education and the promotion of evidence based practices, both in government policy and elsewhere in New Zealand public life. We believe that all government policy decisions should be informed by the best available evidence from relevant groups. We believe that legislation should be based in reality and evidence, rather than on unproven ideas.
 

The Proposals


Our main concern is regarding Proposal five, with the addition of incitement to discriminate to the Human Rights Act.

We accept that inciting others to hate a group of people for their religious belief can be (and has been) a problem in this country, and that it makes sense to have some level of legislation to protect people from this behaviour. 


We also acknowledge that there is a public interest in protecting people from incitement to commit acts of violence and/or to cause harm to others, such as in the case of terrorism.

We believe that the proposed new offences of incitement to discriminate and incitement to hatred are problematic for several reasons, and we are writing this submission to highlight the problems associated with these proposed changes.

Firstly, the proposed offences are deeply problematic because of their language. The wording of the proposal appears to be based on the "hate speech" model, which has been shown to be legally flawed in countries such as the United States, Canada, Australia and the United Kingdom. The "hate speech" model of lawmaking is based on the idea that "hate speech" is a form of dangerous speech, and therefore should be regulated as a form of dangerous speech. This assumption is based on the idea that, when someone says something that is "hate speech", it causes people to act on their feelings and to commit violence. It has been shown, however, that this is not true in many cases. Research shows that most people in most countries who commit violence have a different mental disorder (such as schizophrenia) and are more likely to have been victims of violence. It is also shown that people who commit violence are more likely to be victims of violence. The idea that "hate speech" causes violence has also been shown to be false.

To add insult to injury, the proposed offences are vague. They leave too many questions unanswered. What does "discriminate" actually mean? What does "hate" mean? What does "harm" mean? What if someone is just saying something that they disagree with? What if a large group of people are just saying something that they disagree with? How would the "discriminate" be proven? How would the "hate" be proven? How would the "harm" be proven?

As an example of the vagueness, consider the following clause from the proposed Human Rights Act amendment:

(4) Without prejudice to the generality of subsection (3), it is unlawful for a person to incite any other person to discriminate against a person or a group of persons or to incite a person to commit any act of violence or to do any act to cause harm to a person or to a group of persons in the context of a public discussion, debate or other public event, or in any other public place or context.

If someone is inciting a large group of people to discriminate against a group of people, such as race or gender discrimination, is that person committing an offence?

If someone is inciting a group of people to cause harm to a group of people, such as to murder a group of people or to burn down a church, is that person committing an offence?

There are many, many questions about the new offences, and it is not possible to answer these questions from the description. The wording of the proposal needs to be changed to make it clear what is meant by the language "incite to discriminate" and "incite to hate".

Secondly, the proposed new offences are very broad. The Human Rights Act is very broad, and the proposed amendments add little to the breadth of what is already covered. They appear to be largely drafted for the purpose of adding new offences. They do not appear to be drafted to make the law more effective. It would be impossible to draft a bill that is not overly broad.

The proposed Human Rights Act is already quite broad. It bans a number of things that are arguably not discriminatory or hateful, such as incitement to boycott. Even if the Human Rights Act is broadened, as proposed, it would still need to be kept within the bounds of the Constitution.

The proposed offences are also vague. The definition of "hate" (which is not defined) and "harm" (which is not defined) is incredibly broad. Many things that are not discriminatory or hateful can be described as "harmful". To use an example from the Bible, the book of Isaiah says that Yahweh is "a jealous God and an avenger, who takes vengeance on his adversaries." This has been used by many people as an example of Yahweh being hateful towards other religions. However, it is not hate speech. It does not incite violence or discrimination against other religions. It simply tells the truth about Yahweh. The book of Isaiah also says that Yahweh will judge his enemies, and that Yahweh will be "a refuge for the oppressed and a refuge in times of trouble". This is not hate speech. It is stating the truth about Yahweh. 

We believe that the wording of the new offences needs to be significantly changed to make them more clear. In particular, we believe that the words "hate" and "harm" need to be clearly defined.

Thirdly, the proposed new offences appear to be based on the assumption that people can be "protected" from certain ideas. This is an unproven theory. There is no evidence that people can be protected from certain ideas, and this evidence is being ignored.
 

QAnonsense

Using an older algorithm, GPT-2, with a fairly small data set that it had been pre-trained on, I fine tuned this particular piece of software on the entire back catalogue of QAnon posts. Below are 6 QAnon posts - three are the genuine article, and three are fakes created by GPT-2. Can you figure out which is which?
 

#FactsMatter - Truth to Power - Facts = Power of OP.
Who controls the truth?
Who controls the narrative? #FactsMatter #FactsMatter
We know.
Q

 

>>2434941
Think WRAY (yesterday).
Think [RR] (today).
Comms understood?
Do you believe in coincidences?
Q

 

What happens when the public finds out the TRUTH?
What happens when the FAKE NEWS MEDIA can no longer control the narrative (when disinformation is no longer swallowed and/or accepted as auto-truth)?
CHANGE IS COMING.
THE GREAT AWAKENING.
Q

 

>>533922
Think for yourself.
I know you can 'defend your position' if you can be 'truth-telling' and 'vocal' about what you feel.
But this isn't 'public opinion.'
These people (or 'others') are 'terrorists,' and this will 'kill them.'
We were scared for you. 
You are needed.
Q

 

>>7087382
Note the date of the post - Nov 2017.
Note events happening today.
Reconcile.
News unlocks.
Q

 

Will the MSM push the lie re: Russian collusion?
Prevent if necessary.
Will they have the courage to conduct investigation if found to be Falsifiable? 
Will they have the decency or the ability to report true facts re: collusion?
Will they have the courage to conduct an unbiased investigation if found to have been Falsifiable?
Prevent if necessary.
The Great Awakening.
Q

 


If you've played along, and want to know the answer, the first, fourth and sixth are made by the AI, and the second, third and fifth are genuine QAnon posts. I have a project planned for these AI generate posts which I'm hoping will be a little bit of a fun game that shows just how silly QAnon's posts are. I'll let you all know as soon as it goes live.


 

Trying to create a new Religion


This one surprised me a little. I fed GPT-2 about a dozen holy texts, with the intention of generating some text that was their distilled, combined wisdom. Instead, what I received each time I ran the generator was an attempt to recreate text from just one of the books I'd trained it on. The results are impressive - although the text generated by deep learning doesn't always make a huge amount of sense, I think it could be reasonably argued that the same is the case for genuine holy texts!

This is another game you can play along with at home. All of the texts below were generated by the GPT-2 deep learning algorithm - for each of them, can you figure out which holy text they're emulating? I can't be sure, but I've taken a good guess and included my answers underneath
 

1 Esdr 15:10
Moreover the Lord said unto Moses, If ye speak to any people to speak good, say, Ye shall not speak evil; for they are of an unlawful speech of their own accord.
1 Esdr 15:11
And after the lawgiver said unto the LORD, Speak unto the heathen of the Lord, that they may teach in their own lands of the good word, and that they may not speak wickedly; for they shall never be accepted into their lands of their inheritance, until the day of peace be fulfilled.

 

3:005:008 Thou art the King's counsellor, and I the King his counsellor, and thy word is truth, and I the King's counsellor, and thy word is justice; but the words of the wicked and the wicked are in your mouth: and thou shalt not deliver them, and I the King will deliver thee: so tell me the words of the wicked, as the wickedness of my soul: and thou shalt say, If, when thou art king, thou keepest the commandments of my conscience: then thou art my brother, and I shall not be ashamed.
 

20:2 And the people of Judah went down to the place, and went forth to fight against the people of Jerusalem. Nevertheless, before they did so, they did it with great courage; but not of the same courage was they like unto the city of the children of Israel.
20:3 And the people of Jerusalem fought not as the people of Judah did: but they were not of the same spirit, because the LORD of hosts dealt a dispute with the people of Judah.
20:4 And the people of this city did not rise up against the people, neither went to Jerusalem with the battle; but they went down, and returned as they went down, and departed, and went their way.
20:5 Thus they were scattered, with their tents; and it was not according to their numbers for that they were so scattered in like manner; but according to their number they went up against them, and did battle.

 

They had seen his face from their youth, and were afraid. He had shown great power on the field; their fear was not so great. They found a place that they could put distance between the brother and his master in a distance that was ten paces. He was able to make his brother a prisoner, to cut off their right hand, and to put their master to the death by hanging. His brother then went with his master to the field, and told the people that there were two men at the door, and that the brothers had been there four or five minutes.
 

It is my advice therefore to make use of all means whereby I am able to make use of my intelligence in this matter. For I feel sure that the problems that lie before thee, though greater than any before, and worse still, than were before, and yet, though less, are all to be expected from the evil and wickedness that are in his sight.
 

To-day science reveals the Universe to be subject to the law of relativity, thus, in the sense of a flat, constant state, and yet that the Universe also changes for various conditions, - and this also indicates that matter, is not to blame for the existence of God. We do not find a continuity of matter in matter or a succession in matter over Matter. All that exists does so in degrees and that the higher degree of being leads to other degrees and that all that takes place rests upon this higher level of being.
 

My guess for what these texts are emulating is:
  1. The Apocrypha
  2. The King James Bible
  3. The Book of Mormon
  4. The Koran
  5. The Kitáb-i-Aqdas
  6. Science and Health With Key to the Scriptures

Any other ideas?

One idea I've been given is to find a large set of Deepak Chopra quotes, and use that to train an algorithm to create nonsense quotes talking about quantum realities and the collapse of the wave function. All I need to do now is find a bunch of quotes in a format I can feed to the algorithm.

Do you have any other ideas of what I could feed one of these algorithms?
If you have any news or thoughts you would like to see published in this newsletter, we would love to hear from you at:
[email protected]
if you want to support us by becoming a financial member, or would like to check your membership status, please go to:
https://skeptics.nz/join
Twitter
Facebook
YouTube
Website
Email
Copyright © 2021 NZ Skeptics, All rights reserved.


Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

Email Marketing Powered by Mailchimp

Recommended Posts