Live data from Hacker News

Human brain compresses working memories into low-res ‘summaries’

directorsblog.nih.gov

111–120 of 349 posts

Re: Human brain compresses working memories into low-res ‘summaries’

#111
post #109

Earlier quoted context omitted.

Bread is defined as anything I think bread is, and the same goes for any other word. To hold another position would be in some way dishonest.

That’s not a definition :) And, by the way, a definition is not defined as whatever one thinks a definition is.

Maybe their definition of definition is your definition of bread?

Re: Human brain compresses working memories into low-res ‘summaries’

#112
post #109

Earlier quoted context omitted.

That’s not a definition :) And, by the way, a definition is not defined as whatever one thinks a definition is.

Maybe their definition of definition is your definition of bread?

Give us this day our daily definition

Re: Human brain compresses working memories into low-res ‘summaries’

#113
post #51
post #35

Compression is a component of general intelligence. A few years ago I was very sceptical of machine learning ever leading to general intelligence. I've since changed my mind. There are a lot of parallels to this work and the concept of "embeddings" in machine learning. Intelligence requires the ability to generalize. A prerequisite for generalization is the ability to take something high-dimensional and reduce it to…

The human brain also forgets, something that may be a feature instead of a bug. Also, beyond compression––brains are simulation machines: imagining new scenarios. Curios to understand if ML provides anything analogous to simulation that isn't rote interpolation.

I am quite a novice in ML topics, but isn’t this concept of simultaneously training a generator and validator sort of this?

I don’t know the exact term but I think of deep fake generators with an accompanying deep fake recognizer working in tandem bettering each other constantly?

Re: Human brain compresses working memories into low-res ‘summaries’

#114

Earlier quoted context omitted.

Maybe their definition of definition is your definition of bread?

Give us this day our daily definition

At risk of really devolving this thread, I’m pretty sure that bodybuilders generally agree that bread is counter-productive in the pursuit of definition :)

Re: Human brain compresses working memories into low-res ‘summaries’

#115

It all works this way (though a certain female family member would disagree, claiming to remember conversations word-for-word years later). But my memory works this way. A summary "party at so-and-so's house, weather was nice, overall vibe was ___". The rest is context. You know what the house/backyard is like, you know the general feel of that time of year, you know the crowd that usually comes, you can easily synth…

I think some memories are closer to lossless compressions than lossy which I wonder if it's more of a scale where memories can slide between the two modes with varying degrees of fidelity. Like there are memories that I know I shouldn't remember from childhood that I can remember clearly and others I barely remember what year it happened. So I have to wonder if some of this seeming lossless-ness is more fractal-like…

I vaguely remember reading something that traumatic or “very important” memories never go through the usual process of becoming memories. Instead, when you recall them, your brain physically “relives” it so it is never forgotten. Probably a evolutionary trait to make sure we learn as much as we can from the experience. This is also why you remember those “times you almost died” in slow motion. Your brain goes into a high resolution mode in those cases, which you remember as slow motion, like speeding up a camera and playing it back at normal speed.

Sorry I don’t have any sources, I’m just a casual reader in this space.

Re: Human brain compresses working memories into low-res ‘summaries’

#116
I definitely construct scenes from a few noted details plus general context. Like what colour is my neighbour’s front door? Not sure, even though I pass it every day.

However if I mentally retrace my steps within a short timespan, it seems that I recall details that I would generally not remember. For instance if I leave my house and think, “Did I brush my teeth?”, I can usually confirm/disconfirm by picturing something very specific like where I placed the toothbrush afterwards.

Re: Human brain compresses working memories into low-res ‘summaries’

#117
post #102

I think this is why it's hard sometimes to argue in support of something you believe, even if you're right. At one point, all of the relevant facts and figures were loaded into your working memory, and with that information you arrived at a conclusion. Your brain, however, no longer needs those facts and figures; you've gotten what you needed from them, and they can be kicked out of working memory. What you store the…

One way to make this clear to yourself is to observe how much more difficult it is to "define bread" than it is to answer "is this bread?"

There can be surprising insights yielded from such an exercise. For example, if I think about what separates breads from cakes and muffins, I am forced to deal with the way that a typical "banana bread" (baked with lots of sugar and without yeast) is really a bread-shaped muffin more than a banana-flavored bread. This might seem overly semantic, but it does reflect differences in how it is baked and what it means nutritionally.

Re: Human brain compresses working memories into low-res ‘summaries’

#118

Earlier quoted context omitted.

A nice definition of intelligence I've heard is exactly the ability to form models of the world with predictive power. And a model is essentially a compression of real-world data. Physical laws are a great example of this.

How do you tell if something you're trying to determine as intelligent or not has formed a model?

If it efficiently ingests data with a non-trivial signal-to-noise ratio and returns actions/reactions that contain more signal and less noise.

Re: Human brain compresses working memories into low-res ‘summaries’

#119
> The new study, from Clayton Curtis and Yuna Kwak, New York University, New York, builds upon a known fundamental aspect of working memory. Many years ago, it was determined that the human brain tends to recode visual information. For instance, if passed a 10-digit phone number on a card, the visual information gets recoded and stored in the brain as the sounds of the numbers being read aloud.

I'd be cautious over-generalizing that result, because I think it's also been found that different people do this in different ways, and it may be one of the things that distinguishes speed-readers from other readers.

I know when I read text, my brain sounds it out. It's gotten very fast at it, so I can read pretty quickly, but that sounding-out engages auditory parts of my brain that make it hard to read and listen to someone at the same time. Other people I've met simply do not have that limitation, and their description of the qualia of how they read doesn't mention a sounding-out step at all.

Re: Human brain compresses working memories into low-res ‘summaries’

#120
post #102

I think this is why it's hard sometimes to argue in support of something you believe, even if you're right. At one point, all of the relevant facts and figures were loaded into your working memory, and with that information you arrived at a conclusion. Your brain, however, no longer needs those facts and figures; you've gotten what you needed from them, and they can be kicked out of working memory. What you store the…

One way to make this clear to yourself is to observe how much more difficult it is to "define bread" than it is to answer "is this bread?"

This thought experiment ends with Diogenes running into the Academy and tossing a Guinness in my face. ;)
Post reply on HN