So, but on the lowest level, um, for instance, neurons try to predict the patterns in their environment and react to them. And so, the way we perceive reality, I think, can be understood like a synthesizer. Okay. So, for instance, when you get signals coming in from your

cochlea, you don't take the frequencies of sound because neurons cannot oscillate at 20 kilohertz, and neurons like to oscillate at something like 20 hertz. So, you first decompose the sound into different energy spectra and you have little hairs

in this snail-like organ in your inner ear, and they react depending on the energy of the sound at different frequencies and they send this to the brain. And the brain is now trying to make sense of these patterns, and it makes sense of

these patterns by being able to predict them. So, when the sound starts, it tries to continue the sound to know what's happening next, and when it does this, it's able to explain it. So, you can do this very similar to a synthesizer. You have a few knobs and you

have an oscillator, and there's some almost random connection between circuits, and you tune them until you get the sound you want. And when you are used to your particular synthesizer, you get pretty good at this, right? And you don't only do this for sound, but you can also

do this for vision. So, you look for spatial frequencies, for colors, and so on. And you build lots and lots of synthesizers in your brain that are able to predict and continue these patterns once they start occurring, and once they are done with training and they don't get

better you combine the synthesizers and try to find the meta patterns, the patterns in the patterns, and across different modalities and so you find for instance percepts, you realize that different sounds are related by pitch, you can explain a family of sounds by just changing the

pitch between them or the loudness and ultimately you get something where you unify all the simulations and all the percepts in your mind and what you get is a cohesive simulation where everything moves around in three dimensional space because it's the

simplest explanation that we can find for the things that are going on in our cochlea and field of vision and it's basically the biggest unsolved task in AI, how to do this unified model of the world instead of having a bunch of classifiers that react to individual

feature complexes in the world, learn everything into a single function, into one universe and this is what we're able to do right? You you try to make sense of the universe in one function, there's one way in which reality works. Yeah, yeah, yeah. In some sense the

because the biggest step in the maturation of a mind is understanding its own nature, it's understanding how we work and AI is the best chance that we have in finding it out.

The way we perceive reality can be understood like a synthesizer - @SimulationSeries

Joscha Bach Bits

종합으로서의 신경 예측 및 인식

뉴런은 환경 패턴을 예측하고 그에 따라 반응합니다.

소리 인식은 달팽이관의 유모세포가 소리를 에너지 스펙트럼으로 분해하여 처리하는 과정을 거칩니다.

뇌는 패턴의 연속성을 예측하여 패턴을 해석합니다.

이 과정은 신디사이저를 사용하여 원하는 소리를 내기 위해 매개변수를 조정하는 것과 유사합니다.

이러한 예측 합성은 시각에도 적용되어 공간 주파수와 색상을 처리합니다.

초기 훈련 후 신디사이저를 결합하여 메타 패턴을 식별합니다.

이를 통해 음높이에 따른 소리와 같은 지각 간의 관계를 인식할 수 있습니다.

통합 시뮬레이션을 통해 모든 것이 3차원 공간에서 움직이는 응집력 있는 모델을 만들 수 있습니다.

별도의 분류기가 있는 현재 시스템과 달리, AI에서 통합된 세계 모델을 만드는 것은 큰 과제입니다.

인간은 하나의 기능을 통해 현실에 대한 통합적인 이해를 달성합니다.

마음의 본질을 이해하는 것은 마음의 성숙을 위한 중요한 단계이며, AI는 이를 달성할 수 있는 최고의 기회를 제공합니다.

The way we perceive reality can be understood like a synthesizer - @SimulationSeries

The way we perceive reality can be understood like a synthesizer - @SimulationSeries

스크립트