Can AI learn to understand humor.

[Write4U]

I think this may turn out to be a very interesting test of AI ability for abstract thought.

That's funny -- but AI models don't get the joke
Date:July 31, 2023
Source: Cornell University

Large neural networks, a form of artificial intelligence, can generate thousands of jokes along the lines of "Why did the chicken cross the road?" But do they understand why they're funny?

Using hundreds of entries from the New Yorker magazine's Cartoon Caption Contest as a testbed, researchers challenged AI models and humans with three tasks: matching a joke to a cartoon; identifying a winning caption; and explaining why a winning caption is funny.

https://www.sciencedaily.com/releases/2023/07/230731122233.htm

To understand a joke you need a general understanding of what is a normal experience and what about abnormalcy can be funny. This is usually associated with biology . But can AI learn to make a distinction between normal and different in a "funny" way?

 

[Write4U]

Cause AI its the same a google or other search services.

No, AI is more than that. With complex systems, there are emergent abilities that are over and above the sum of the parts. Intelligence (biological or artificial) is such an emergent ability.
The GPT series AI can think via a feedback system that kinda resembles human thought processes.
GPT AI can think!
Check out Leta and the GPT4 who probes her record.

 

[C C]

I think this may turn out to be a very interesting test of AI ability for abstract thought.

That's funny -- but AI models don't get the joke
Date:July 31, 2023
Source: Cornell University


https://www.sciencedaily.com/releases/2023/07/230731122233.htm

To understand a joke you need a general understanding of what is a normal experience and what about abnormalcy can be funny. This is usually associated with biology . But can AI learn to make a distinction between normal and different in a "funny" way?

Even many humans might not be able to conceptually explain why some events or stories are funny. In those cases, the laughter or whatever response is instinctive and any knowledge or understanding of "why" is tacit.

AI may be generating jokes from one of its emulative, statistically-guided block boxes -- without the definitive rules/principles, templates, etc that humans use for demystification.

There are categories of humor, theories of humor, philosophy of humor, etc. That is, if we can covert humor to explicit knowledge -- precisely articulate why something is funny and classify it and explain why it belongs to _X_ type, and codify such as instructions... Then AI has what it needs to explain or understand why _X_ was funny.

Actual laughter or an inner feeling of mirth is unnecessary. Since physical description as well as the procedural operations of computer hardware do not capture and carry the qualitative characteristics and manifestness of human experiences, anyway: feelings, visual images, odors, etc.

Even if machines did have private states, they can't be accessed and vetted. The same would be the case with "other humans" if each individual didn't have their own inner proof and be forced to morally and via consistency attribute the same capacity to the rest of the bipedal herd.
_

 

[Write4U]

I believe it has something to do with the listener's expectation built by an elaborate story that leads to a completely unexpected closing "punch line" and shocks the listener's sense of logical conclusion.

The laughter is the cognition and admission of having been fooled (trapped) into a totally unexpected turn of events.

I wonder what an Chat AI would make of this little skit by Victor Borge, that contains many subtle variations of humor.

Can AI watch TV?

 

[thetechload]

Yes, AI can learn to understand certain aspects of humor, but it’s a challenging process! Humor often involves complex language play, cultural context, emotions, and an understanding of human experiences, all of which can be tricky for AI to fully grasp. Here are a few ways AI approaches humor:

1. Pattern Recognition: AI can learn to recognize basic structures of jokes (like puns, wordplay, or unexpected twists) by analyzing large datasets of humorous content. This is why some chatbots or apps can produce jokes that are funny, although they may sound repetitive or formulaic.
2. Sentiment and Context: Understanding sentiment (like sarcasm or exaggeration) is part of humor, and AI models have been trained to detect these subtleties with some success. For example, models that analyze social media posts for sarcasm look at language patterns, context, and even emojis to interpret what’s intended as humorous.
3. Cultural References and Context: Humor is often deeply cultural. An AI trained on Western humor might miss references or humor styles from other cultures, but with enough exposure to diverse examples, AI can begin to catch on to cultural references and styles.
4. Timing and Delivery: Timing is crucial in humor. Comedic timing is incredibly hard to program but can be simulated to an extent, especially in spoken or text-based AI that is trained to avoid overly long pauses or abrupt responses.

While AI is improving at generating jokes and even responding with humor, there’s still a long way to go before it can match the nuance of human wit. Most current AIs handle humor by learning common structures and responses but still struggle with subtleties like irony or social cues in real-world interactions.

 

[wegs]

Sure, humor relies on cultural context, wordplay, ''punchline'' timing, and shared experiences, which I'm assuming would be (currently) challenging for AI. But not impossible. AI can respond to humorous prompts (like any other prompt), however, the subtleties that make something genuinely funny might not come through. That said, how we all interpret jokes and humor, is subjective. Maybe someone will find AI's humor to be utterly hysterical while some of us may not see any humor in it at all. Not because the jokes or tone isn't funny, but we may not find it funny. That's the thing about humor, it's subjective. Will AI become sentient one day? Will we be able to know the difference?

 

[Write4U]

It should be possible with a feedback system that is programmed to anticipate the next set of words as humans do.
When it does not arrive, in Humor it is called "misdirection". To an AI it becomes a "differential equation".

The trick is knowing the difference when a misdirection results in an anticipated answer that is either tragically wrong, or funny wrong.
IMO, those responses can be "learned".

Watch this delightful scene where a young orangutan "gets" the magician's joke. This so funny...

Artificial Intelligence Can Now Craft Original Jokes—And That’s No Laughing Matter

Artificial Intelligence Can Now Craft Original Jokes—And That’s No Laughing Matter

Researchers have trained artificial intelligence to make jokes, but the advances are making many—and not just comedians—uncomfortable.

time.com

Antagonism between brain regions relevant for cognitive control and emotional memory facilitates the generation of humorous ideas​

Antagonism between brain regions relevant for cognitive control and emotional memory facilitates the generation of humorous ideas - Scientific Reports

The ability to generate humor gives rise to positive emotions and thus facilitate the successful resolution of adversity. Although there is consensus that inhibitory processes might be related to broaden the way of thinking, the neural underpinnings of these mechanisms are largely unknown. Here...

www.nature.com

 

[exchemist]

It should be possible with a feedback system that is programmed to anticipate the next set of words as humans do.
When it does not arrive, in Humor it is called "misdirection". To an AI it becomes a "differential equation".

The trick is knowing the difference when a misdirection results in an anticipated answer that is either tragically wrong, or funny wrong.
IMO, those responses can be "learned".

You still have absolutely no idea what a differential equation is. A differential equation is one in which a variable and a derivative of that variable both appear. A simple example is radioactive decay: -dN/dt = λN, where N is the number of undecayed atomic nuclei and λ is the decay constant. N is the variable and dN/dt is the rate of change of N with respect to time, i.e. the first derivative of N with respect to time. What this is saying is that rate of decrease of undecayed nuclei is proportion to the number that remain at any given time.

That is a differential equation. Unless you have something like that, i.e. including a variable and one or more derivatives of it, you do not have a differential equation.

Don't abuse terms you do not understand.

 

[Gawdzilla Sama]

Jokes tend to be formulaic, so an AI could be trained to spot that standard formulas and do a sensitivity check on the joke before repeating it.

 

[Write4U]

You still have absolutely no idea what a differential equation is.

You are still speaking subjectively in Human mathematics. The Universal mechanics do not "think" in human terms.

i.e. the first derivative of N with respect to time. What this is saying is that rate of decrease of undecayed nuclei is proportion to the number that remain at any given time.

Differential forms are totally anti-symmetric tensors. A joke is an abstract asymmetrical object or function based on a dynamic pattern of mathematical differential equations.

A joke is a dynamic narrative (tensor), building an expected anticipation in the listener and the humorous self-deprecating cognition (release) in the unexpected expressed differential expression. This dynamical sequence can lead to some very interesting universal results.

Numerical relativity

Numerical relativity is the sub-field of general relativity which seeks to solve Einstein's equations through the use of numerical methods. Finite difference, finite element and pseudo-spectral methods are used to approximate the solution to the partial differential equations which arise. Novel techniques developed by numerical relativity include the excision method and the puncture method for dealing with the singularities arising in black hole spacetimes. Common research topics include black holes and neutron stars.

Perturbation methods

Main article: Perturbation methods in general relativity

The nonlinearity of the Einstein field equations often leads one to consider approximation methods in solving them. For example, an important approach is to linearise the field equations. Techniques from perturbation theory find ample application in such areas.
This section needs expansion. You can help by adding to it. (June 2008)

The nonlinearity of the Einstein field equations often leads one to consider approximation methods in solving them. For example, an important approach is to linearise the field equations. Techniques from perturbation theory find ample application in such areas.

Mathematics of general relativity - Wikipedia

en.wikipedia.org

Tensor-release mechanisms in humor is the establishment of a believable narrative for an entertaining asymmetrical situation which concludes with a "punch line" the release of the anticipated answers in a totally unexpected direction.

As the brain is capable of calculation, it follows the mathematical guiding principles involved in the "mathematically humorous formula"

 

[exchemist]

You are still speaking subjectively in Human mathematics. The Universal mechanics do not "think" in human terms.

Differential forms are totally anti-symmetric tensors. A joke is an abstract asymmetrical object or function based on a dynamic pattern of mathematical differential equations.

A joke is a dynamic narrative (tensor), building an expected anticipation in the listener and the humorous self-deprecating cognition (release) in the unexpected expressed differential expression. This dynamical sequence can lead to some very interesting universal results.

Numerical relativity


Perturbation methods

Main article: Perturbation methods in general relativity

Mathematics of general relativity - Wikipedia

en.wikipedia.org

Tensor-release mechanisms in humor is the establishment of a believable narrative for an entertaining asymmetrical situation which concludes with a "punch line" the release of the anticipated answers in a totally unexpected direction.

As the brain is capable of calculation, it follows the mathematical guiding principles involved in the "mathematically humorous formula"

View attachment 6239

Nonsensical gobbledegook from start to finish.

 

[Write4U]

Nonsensical gobbledegook from start to finish.

The LOL Formula: How to Mathematically be Funny​

We’re working on a comedy web series together, somewhere on our second carafe of coffee, and we start talking about comedy. The dude gabs a pen, and he sketches on the back of a napkin his interpretation of what makes something funny. It was a Venn diagram with the word “normal” in one circle, “absurd” in the other circle, and “comedy” in the meeting space.

I almost start laughing at this and tell him, “hey buddy, I’ll handle the funny,” but then it hits me: he’s right.

It’s something we had been talking about in improv for a long time, using phrases like “finding the game” and “grounding,” but never had I seen it exposed in such a simple, naked way. The formula for comedy:

Normal + Absurd = Funny

The LOL Formula: How to Mathematically be Funny — Cideshow

For eight years now I’ve taught comedy improv to engineers, coders, and other people who are not often stereotyped for being riots...

cideshow.com

It's a differential equation, no?