Dieser Bereich kann Inhalte enthalten, die nicht für alle Nutzer geeignet sind. Dazu können unter anderem Texte, Medien oder Diskussionen gehören, die als beleidigend, extremistisch, gewaltbezogen oder anderweitig belastend empfunden werden. Wenn du solche Inhalte nicht sehen möchtest, nutze bitte die jeweiligen Filter- und Meldeoptionen der Plattform oder meide entsprechende Threads/Communities.
the greatest power of machine learning algorithms is the source of its greatest drawback. they are essentially heuristic models of something, constructed in a way that they are much cheaper to execute than a traditional algorithmic approach. this cheapness results in error, which for a lot of applications is fine because you can refine/check the result with more accurate tools, but it also means you can never just trust it like you can with more traditional tools. this problem is baked into the technology so theres no amount of scale that will make it go away, as we’ve seen time and again with LLMs. failing to understand this is the mistake most people make when it comes to “AI”, and results in all kinds of bad decision making. but in the hands of people who understand the limitations, machine learning can truly be a game changing technology. not chatbots though those are fucking stupid and i hope they go away when the bubble bursts.
I disagree with your assessment. Modern machine learning methods will result in a far more complex algorithm than any human could device on their own. It will be “cheap” to execute, but only because of the vast amount of data it’s been trained on. That’s the strength. The raw number of calculations it can do per second means it can sift through more data in a day than a human could in its lifetime.
The biggest danger is hidden local minimums that can potentially distort new results and findings. But that’s not unique to machine learning. And you can let the AI grade its result with how certain it is. Automatically flag any result that falls below a certain threshold. And again, this is where the danger of local minimum come in. Because it can raise the certainty of results when maybe it shouldn’t be so certain. Though you will find out quickly when results are false with high certainty. Then you go back and you keep working at it.
As of this particular topic of earthquakes. There are no traditional tools of accurately predicting them. We’ve never been able to do it before. Now they think this might be able to. So we will just have to wait and see if its predictions are accurate or not.
i can see that youve had some coolaid. i hope you can stop drinking and start recovering. a lot of what you said is gibberish, but throwing more data at machine learning models has been shown to result in diminishing returns, and doesnt change that the technology is inherently fallible. it might get very close to 100% reliable, but the hallucination problem will never fully go away. thats not the end of the world, but it requires machine learning algorithms to be paired with more deterministic tools to be sure of results. it happens with our own brains too, which is part of why we invented math and computers in the first place.