My definition of "understanding": In this universe, there is things/ problems/ states, and you can't "know" how to solve them i.e. what will happen after thing_1 (ex. 1>2/3/4), and pick which future you want (1>69), unless you are dealing with patterns. So to solve a state is to know what comes next, if you can't predict, then you can use past experience, hence can't solve/ predict/foresee. In this universe, we can, because there is patterns. We can use past experiences to inform us of what should happen next for some state or states we could try, whether it is trying to exit a fire or extending a cup design (where to put the handle, what should come next).
So our universe has patterns that reoccur, and we can only use this to solve problems. We can only use patterns, or randomness, to solve problems. You may randomly move to avoid a bullet, this is randomness, but you are trying to stay a pattern (survival), by using patterns (memories and house/world). Acting random on purpose is you using patterns/memories to predict what to do/ what will happen next. So to understand past experiences/ data, you want to find patterns. These can be exact matches, or delay and/or holed matches ex. here are 2 similar memories: 123456 = 12456. Another: 123456 = 12h456. Anther: 123456 = 1h2h3h4h5h6. Another way to match a memory experience is translation by using side matches (which can be similar) ex. dog ate = cat ate.....so dog = cat because their side contexts are similar. This allows you to see dog and predict meow comes next. This actually helps. So does using hole/delay matches. And you probably just read "so dog" and realized I said "so does" actually right? I did. Counting if dog>meow is more probable than dog>sleep also helps, it happens more after that word maybe if look through the text data, you may see dog sleep less often! Also checking recent history ex. cat cat cat cat cat, what word comes next? So recency has a big impact too! Patterns stick together, a topic/ article/ movie will likely be on a single subject lots ex. all about dogs or space ships. When you see "the cat and dog are here, and it is so cute!", the 'it' refers to a rare word/pattern in this sentence (but not too rare), is it dog or cat it refers to? You could use the fact that most people buy dogs (if they do, idk), or that dog translates to ugly (not to me or others I'd say though). And note that seeing dog....then it in the sentence is the recency thing; seeing cat 10 times and then ca>? predicts strongly a 't' comes next, so seeing 'cat.....it', the it is trying to match recent history to predict what came next after this context recently, so it can be lots of words, see? So we are pattern matching here, cat = cat/dog/it is this sentence. So, understanding, is the looking for patterns. What is life/alive is not aliens but patterns that last long or come in large quantities. We use patterns (memories/ homeworld) to be patterns (survival). Knowing where everything is in your world leads to better prediction, you can walk through your home blindfolded using less energy to find things, if all is the same shape (square/ circle), and grouped (restaurants, construction, clothes), and lined up and timed together. The future world will be a 3D gravityless airless metaloid-nanobot sphere that can instantly regenerate and morph into tools as it desires, a fractal. We only like random paintings etc when young because we explore for new data and not a specific domain, exploring is needed because things are hot like the big bang and it is trying to find patterns to cool down and last. DNAs form mom and dad merge like memories trying mom and dad traits ex. wings+man= birdman, or wheels+home= trailer, this tries new ideas, but is not as smart as brains doing it. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tf91e2eafa2515120-Mdffd67c213faf4f46b9f8c85 Delivery options: https://agi.topicbox.com/groups/agi/subscription
