[zaptheimpaler]

The best theory Ive read is the theory of constructed emotion by Lisa Feldman Barrett. It directly contradicts this theory. Emotions are basically very fast predictions by the brain. The whole idea of emotions being completely independent from cognition just seems pretty dated today.

[andirk]

Isn't trauma often permanent? Even if it's that a kid got his red fire truck stolen from him in kindergarten and now he has attachment issues. Or someone deals with horrible things in their past. It can be ridiculous or completely founded, but it's not fleeting or "very fast". It's not DNA but it _seems_ engrained, however incorrect the conclusion may be.

[SamoyedFurFluff]

Trauma can be lifelong if it isn’t worked through, but some trauma can be worked through (and hallucinogens are showing promise of working through trauma quicker than medication&talk therapy alone). I think the actual emotional trauma response however is individually maybe 10-15 mins long at maximum, but feeling the emotion often causes you to think other triggering thoughts or behave in a manner exacerbating the trigger, thus making an individual 10-15 minute thing take hours to sort itself out.

[amelius]

> The best theory Ive read

You can't judge a theory just by reading about it.

[ryanklee]

Yes, you can.

If the theory is well-formulated, rests on vetted facts, and you understand the context sufficiently, you can say: "This is a good theory." If it's not well-formulated or does not rest vetted facts, and your understanding is sufficent, you can say: "Bad theory."

There's an entire world of variation between those 2 extremes, but there's no reason implicitly why you can't understand a theory just by reading about it.

[amelius]

Only true if collecting "vetted facts" is considered to be just reading and not doing hard, real work.

[ryanklee]

Totally wrong.

If the goal is to evaluate a theory's virtues, there's nothing stopping you if you have access to the correct documentation (provided it exists). Not sure what you think is missing.

[amelius]

We're talking about new theories here, not well-accepted theories.

[ryanklee]

Not correct.

"You can't judge a theory just by reading about it."

Also, can't see why it should matter one way or the other.

[ryanklee]

Not correct.

"You can't judge a theory just by reading about it."

[flangola7]

Give examples

[treeeman]

Not a dementi, rather an open question I have: How does the constructed emotion theory contradict the idea that emotion is generated as a response to observed "behavior" and other sensual input such as own and other's body language? (As opposed to the believe that a certain emotion induces specific behavior.) Cognition and emotion would only be separable if body language and other relevant stimuli (which are believed to be largely outside conscious control) could be turned off, to prevent the generation of a fitting emotion - so the separation seems to be mostly conceptual/hypothetical.

[thenerdhead]

She builds upon these theories as her main argument is that Descartes, Spinoza, Malebranche, and James were mostly "fixed" perceptions of physical states. Hers includes more "dynamic" means like sensory input, past experiences, culture/locale, and contextual factors.

But I don't see that. When I read James or Spinoza, I consider the same factors she includes from her book. They all generally believe in the bi-directional power of the mind-body connection and how that relates to someone's driving philosophical system.

[barberpole]

> Emotions are basically very fast predictions by the brain

This predicts that basically ChatGPT has emotions.

[FrustratedMonky]

and why not? Both are neural nets. Carbon versus Silicon.

It's just scale at this point. At this point, maybe human neurons and connections are more complex than what can be done in a machine, but not for long.

Humans only think they are special, that a machine can't feel.

But that is only our subjective experience of ourselves, which itself is also just interpretation. It seems like everyone is in agreement that our perception of the outer world is fallible. But so is our perception of our internal processes. Our inner thoughts are just as opaque to us as the outer world.

"Man can do what he wills but he cannot will what he wills." Schopenhauer.

[adammarples]

Neural nets and brains are fundamentally different. If you cut open a brain you won't find attention blocks. There's no evidence as far as I know that back propagation actually happens. It's a nice model but it is a model.

[FrustratedMonky]

I think the latest neuroscience would disagree.

Maybe if you cut open a brain, you wont find a printout with backpropagate code.

But you do find neurons, in a network.

Humans do learn right? They take in information and encode it in their brain.

Why does it have to be back propagation to qualify. Bayesian? Minimum Entropy?

There are a lot of forms a machine neural network can take. There are a lot of theories on exactly how the brain 'calculates'/'processes'. With all of the advancements in Neuroscience and AI in last 5 years, it is bit hubris to say we'll never be able to figure out the brain, and also be able to model it.

There are a lot more like this. The field is moving too rapidly for me to go find every paper today. But it is dozens, and not even so cutting edge there isn't already books on it.

https://www.quantamagazine.org/some-neural-networks-learn-la...

[kelseyfrog]

To add, there is a whole field of biologically-plausible learning. Predictive coding is one such method. It is isomorphic to SGD[1], and doesn't require global computation. ie it's backprop-free optimization. Although slower than SGD on silicon, it's massively parallelizible which ends up making it more biologically plausible rather than less.

1. 2020 PREDICTIVE CODING APPROXIMATES BACKPROP ALONG ARBITRARY COMPUTATION GRAPHS https://arxiv.org/pdf/2006.04182.pdf

[paint]

Because human cognition is fundamentally affected and controlled by it's body and its sensorimotor interactions with the environment (read: "Six Views on Embodied Cognition" by Margret Wilson) which LLMs don't have and will never have.

[FrustratedMonky]

1

Agree with embodied issue. Humans have a large amount of sensory input, from the 'body'. I'd disagree that AI will never have this, considering the large number of sensory technologies that already exists and are being developed.

It only takes wiring them together. Eyes, smells, touch, these are all existing and being refined.

2.

LLMs? AI research is far more vast than just LLM's. LLM's just happen to be the latest shiny thing.

"never"?. That is bold, 5 years ago people said the abilities in current LLM's were a "never", yet here we are.

[thfuran]

What makes it best?

[gushogg-blake]

It's been a while since I read it, but one of the central ideas is that we construct instances of emotions based on sensory data (from the internal systems that detect and predict our energy needs) in a similar way to how we construct instances of concepts from other sensory modes, e.g. seeing a particular cluster of visual features and concluding "that's a bee". When we feel anger, that's our brain's best guess as to the appropriate interpretation of a bunch of internal measurements -- raised blood pressure, high arousal, etc -- based on context. The same internal state in a different context could be interpreted as a different emotion, so there is no consistent or unique biological fingerprint behind a given emotion.

It's counter-intuitive, but it fits better with what we know about how the nervous system works than the commonly accepted fingerprint idea.

Emotions are the brain's attempt at constructing the concepts that best explain and account for our body's energy budget.

[choxi]

What’s the “fingerprint idea”? That’s not mentioned in the article.

I don’t see why this theory contradicts the one in the article, in your own examples you’re using bodily states as the input to emotional processes.

[gushogg-blake]

I have to admit I didn't actually read the article so wasn't trying to compare it with LFB's model, just wanted to expand on the mention of it. The bodily states are the inputs, but the interpretation and generation of emotions is context- and culture-dependent.

The fingerprint idea is that emotions are caused by body states without that layer of interpretation/construction, so you could look at bare metrics like blood pressure, arousal, and facial expression and derive which emotion the person must be feeling directly from them.

[thenerdhead]

Novelty most likely.