Defeat the Münchhausen trilemma.>pro tip: you can’t
scientific modeling and prediction.
>>17021937pro tip: you are a faggot for using this OP format.
>>17021937Accept that a dogmatic argument isn't inherently flawed. If an axiom is sufficiently obvious then it need not be rigorously proven.
>>17021968>If an axiom is sufficiently obvious then it need not be rigorously proven.two problems with this: axioms are not proven by definition, but chosen. and, it is sufficiently obvious that the earth is flat and proximity to doctors causes diseases.
>>17021937The solution to Munchausen by proxy is to beat women.
>>17022005>axioms are not proven by definition, but chosenCorrect. Which is the whole point.>it is sufficiently obvious that the earth is flatIt would be a fine starting point up until the moment it was disproven.>and proximity to doctors causes diseases.Yeah see how far that axiom gets you.
>>17021937imagine acting so smug about an argument that gets btfo by cogito ergo sum.
>>17022304Imagine posting in a gey thread.
>>17021937Neither Science nor Mathematics depends on deducing things that are unconditionally true. Why does /sci/ need to defeat the Münchhausen trilemma if it's not actually attacking /sci/?
>>17022304>magine acting so smug about an argument that gets btfo by dissociative PTSD
>>17021968>If an axiom is sufficiently obvious then it need not be rigorously proven.Why not?
>>17021937It can be defeated by the autoencoder. In step one, you creare a highly compressed representation of given inputs. Then you take this simplified representation, and use it to reconstruct the original inputs. If the inputs can be accurately reconstructed, it proves that the representation is correct, and sufficient. It isn't hard truth as in philosophy, as new inputs can invalidate the representation, but it's far from the claim that nothing can be known, ever.
>>17022318>nonsensical pseudbabbleAll "AI" fans are mentally ill. All of them. No exception.
>>17021937only if you defeat the CAP theorem first
>>17022323>only if you defeat the CAP theorem firstSaaaar! That is a brilliant argument, saar!
>>17022324thanks batman
>>17022318The theory behind autoencoders in general depends on the axioms of math and logic. To actually model some data using an autoencoder, you then need to make a bunch of explicit assumptions about the data. What you get is an ad hoc, empirical model that makes a bunch of implicit assumptions about the data that you don't even know about. So you have a whole hierarchy of unprovable assumptions here ensuring that your model can always fail for new inputs. You couldn't have failed harder if you tried.
>>17022323Why can't it be defeated like this?:Send empty numbered packets between all nodes. Once a packet is received, the next packet in line is sent back, with a reasonable delay. Updates are added to the packets as needed. No extra confirmation is needed, when the next packet in line is received, it proves that the previous packet was also received. Every N miliseconds, updates sent M packets back are assumed confirmed, and applied. Should follow all criteria as long as the clocks are synchronized.
>Why can't it be defeated like this?:>Send empty numbered packets between all nodes. Once a packet is received, the next packet in line is sent back, with a reasonable delay.>Updates are added to the packets as needed. No extra confirmation is needed, when the next packet in line is received, it proves that the previous packet was also received.>Every N miliseconds, updates sent M packets back are assumed confirmed, and applied.>Should follow all criteria as long as the clocks are synchronized.
>>17022329It doesn't rely on any assumptions, the fact that the original inputs can be reconstructed proves that the model is correct, as the inputs couldn't be reconstructed, if it was wrong or incomplete. The only way it can fail is from the lack of important data.
>>17022339>It doesn't rely on any assumptionsWrong and I've just spelled out how.> the fact that the original inputs can be reconstructed proves that the model is correctWrong. The only way to prove it correct is to test it on every possible future input as well and verify that it's correct.
What is it with the influx of mentally ill, brown "people" spewing random concatenations of CS/ML terms and thinking they're making sense?
>>17022341irony, thy post is (You)
>>17022345>psychotically ill brown insect doesn't understand the words its using
>>17022331>assumed confirmedthis might work if hardware was 100% reliable, but physics unfortunately says>no
>>17022346to this day, i'm unsure if this is a bot or some neet whose deficient brain operates on simple loops. i lean towards bot since i'd rather not live in a world where people this insane exist.
>>17022349>brown insect continues his botlike loop of generic repliesDoesn't matter. >>17022329 stands completely unchallenged.
>>17022340No, you didn't, and I admitted right from the start that you can always find new data that invalidate it. You may be using some impractical definition of truth, that may as well not exist in amy real manner. It may not even be the most reliable approach to to try find the singular truth, you may be better off using multiple models until some of them are conclusively disproven. You may pick the best suited one for a particular application, or run multiple competing models when making critical predictions.Anyway, the autoencoder does provide a way to bootstrap thought that is neither circular, nor infinite regress, nor dogmatic truth.
>>17022353>I admitted right from the start that you can always find new data that invalidate it.Yes, your post was incoherent and self-refuting right from the start. Anyway:The theory behind autoencoders in general depends on the axioms of math and logic. To actually model some data using an autoencoder, you then need to make a bunch of explicit assumptions about the data. What you get is an ad hoc, empirical model that makes a bunch of implicit assumptions about the data that you don't even know about. So you have a whole hierarchy of unprovable assumptions here ensuring that your model can always fail for new inputs. You couldn't have failed harder if you tried.Either way, it's clear that you're brown and nonsentient just from the fact you can't see how your "solution" is just a worse and less reliable variant of normal scientific modeling.On a totally unrelated note: I wonder if any of the "people" """moderating""" this board are brown insects.
>>17022364please solve Hume's problem of induction next
>>17022375>please solve Hume's problem of induction nextYet another mentally ill and incoherent reply.
>>17022364It doesn't depend on any axiom or other assumption, the autoencoder proves its model correct by reconstructing the original inputs from the compressed representation. >>17022375There is no problem of induction, induction is false. Consider the claim that nobody can live to 140, because it's statistically impossible. It's a wholly bizarre claim, because we don't know why people age, and it may very well be possible to live much longer.
>>17022378The theory behind autoencoders in general depends on the axioms of math and logic. To actually model some data using an autoencoder, you then need to make a bunch of explicit assumptions about the data. What you get is an ad hoc, empirical model that makes a bunch of implicit assumptions about the data that you don't even know about. So you have a whole hierarchy of unprovable assumptions here ensuring that your model can always fail for new inputs. You couldn't have failed harder if you tried.>y-y-yes b-b-but it doesn't depend on any axiom or other assumptionYou're psychotically ill.
>>17022379>To actually model some data using an autoencoder, you then need to make a bunch of explicit assumptions about the data.No, you don't. That's why it's called UNSUPERVISED learning. It's just not how it works at all.
>>17022393>That's why it's called UNSUPERVISED learning.An autoencoder is not necessarily UNSUPERVISED learning and UNSUPERVISED learning still assumes things about the data. You're a brown insect and you have no clue what you're talking about whatsoever. You're literally just stringing together terms.
>>17022396What does it assume? What assumptions does it make?
>>17022402>What does it assume? What assumptions does it make?That would depend on what data you're trying to model, you absolute retard.
>>17022375Nta, he needs to be medicated. But I've found this solution. >There are beneficial acts for the individual >There are beneficial acts for the collective >These can be in conflict (see: issues with golden rule)>Kants Categorical Imperative is simply act such that it's beneficial to both Now for the solution to Humes guillotine>Identify what is true for individuals and collectives by observation >Categorical Imperative solves this as what you ought to do Ezpz
>Nta, he needs to be medicated. But I've found this solution.
>>17022433Meds. Now.
>Nta, he needs to be medicated. But I've found this solution.>>There are beneficial acts for the individual >>There are beneficial acts for the collective >>These can be in conflict (see: issues with golden rule)>>Kants Categorical Imperative is simply act such that it's beneficial to both >Now for the solution to Humes guillotine>>Identify what is true for individuals and collectives by observation >>Categorical Imperative solves this as what you ought to do >Ezpz
>>17022437Absolutely seething lmao. Call your mom. She loves you and is worried about you.
>Absolutely seething lmao. Call your mom. She loves you and is worried about you.
>>17022404I was only asking you, because you're obvioisly making it up. No assumptions need to be made
>>17022558>I was only asking you, because ...Because you're too retarded for basic reading comprehension, so when I say "a bunch of explicit assumptions about the data", you obviously can't infer that it depends on the data - that takes an IQ a little above the Indian average.
how does this "person" have the time to reply in 10 mins or less or your money back literally 24/7 lol
>how does this "person" have the time to reply in 10 mins or less or your money back literally 24/7 lol
>>17022313Elaboration on the matter is unnecessary
>Elaboration on the matter is unnecessaryWhy is the brown horde like this?
>>17022624I'm not interested in such philosophical quandaries. You can measure the melanin in my skin and that's good enough for me.
>>17022629>I'm not interested in such philosophical quandaries.Then why are you posting in this thread, brownoid?
>>17022630Specifically because it bothers you and I find that amusing.
>my parents are siblings b-b-b-because it bothers you, ok??
>>17022632Yeah, something like that.
>>17022565It doesn't depend on the data. No assumptions are needed. The fact that you can't provide an example proves that you're making it up.
>>17022922>you can't provide an exampleI just don't know what example to give so that a mentally ill, 80 IQ jeet like you would grasp it. Let's see. k-means clustering is an unsupervised learning algorithm. Applied to data with no distinct clusters, or where clustering isn't a relevant feature of the data-generating process, it accomplishes nothing. When a competent ML engineer uses it, it's because the data is assumed to fall into useful clusters.Did you learn something from this? Are you going to admit your mistake and move on? No. Because you aren't fully human. You're Indian thus you will continue to hallucinate new cope.
>>17023043>. k-means clustering is an unsupervised learning algorithmLol. Lmao even. Nobody tell him.
>>17023052>k-means clustering doesn't heckin' count!!!I expected you to spout some kind of psychotic "retort", but you went for the one refuted immediately by reading any introduction to KMC or just a basic google search.
>>17023056I'm not the anon you originally responded to. I'm a different one who's laughing at you. Believe it or not, multiple people are laughing at you.
>>17023058Ok, I guess India does have approximately 1.5 billion subhumans spamming the white internet nowadays.Here, tell MIT their introductory courses are out of line with Indian IT expertise:https://www.mit.edu/~9.54/fall14/slides/Class13.pdf
It's like saying we are living in the quantum, above where the atom is situated. Which is true.
>Saaaar, we are living in the quantum!
>>17023060>Shits on Indians>Cites Indians Weird strategy. Here I'll give you a hint. I didn't read the argument you had with the other anon. I simply saw that one line in your post, quoted it, and laughed at it. And told nobody to tell you. While I'm tempted to tell you why, I will resolve in leaving you to bask in your ignorance. Here I'll give another hint. I use k-means clustering in my research.
>>17023071>>Cites Indians You're literally psychotic. Nowhere did I cite any Indians.
>>17023073You have ai psychosis btw.
>>17023076>psychotic insect devolves into literally incoherent responses
>>17023077Okay I can see this is getting too abstract for you. I'll break it down. (See how I'm using AI language to appeal to your small amygdala?)I'm not calling the authors of that lecture slides Indian. I'm accusing you of citing AI psychosis. That is, they're propagating the same nonsense that's primarily promulgated by Indians in tech for branding. For someone who hates Indians so much, I think it's important you understand that your argument is citing Indians at its core. Perhaps a Socratic learning approach will work here (pun intended). Can you cite one (1) optimization program that isn't classified as a learning algorithm? Can you then attempt defining the difference between a learning algorithm and one that doesn't learn? Probably not. Fuck it, I'll tell you because it'll be funny. K-means clustering is NOT a learning algorithm. This is branding done by AI tech bros (largely Indian) who are idiots. Your AI psychosis will have you ask AI if k-means is a learning algorithm. It will tell you yes, because it's trained on content written by tech bro idiots.
>>17023082>I'm not calling the authors of that lecture slides Indian. I'm accusing you of citing AI psychosis.You're quite literally losing your mind at this point.
>>17023083>literally losing your mind>literally As in my brain is outside of my skull? Lol. First you don't know the difference between an optimization algorithm and a learning algorithm (because you only are fed information from the AI bros, which is what AI will feed you btw—pay attention to the language used and how it references conventional ML communities when it validates you). Next you don't know what literally means. See the theme? Your imprecision betrays your argument. You're jerking off on your imagined superiority complex over others without realizing you have at best a surface level familiarity with the topic because of your addiction to AI.
>>17023084Imagine thinking I'm gonna read more of your incoherent drivel.See:>>17022329>>17022396>>17023043>>17023060There is nothing further to "discuss".
>>17023088I accept your concession.
>>17023089>i continue to hallucinateDoesn't matter. The undisputed facts of the matter can be found here:>>17022329>>17022396>>17023043>>17023060
>>17023090Spamming your own posts doesn't make them right. Like I said, k-means clustering is a basic optimization algorithm that far predates ML literature and associated, targeted learning algorithms. It's no more a learning algorithm than sorting algorithms (identifying next "token") or newton raphson or rk4 (finding minima). It's only called a learning algorithm by AI tech bros as a grift to get money. Calling k-means a learning algorithm is like calling the median an economics algorithm — it's completely ass backwards. Using a widely applicable mathematical tool (k-means; median) in your discipline (ML; economics) doesn't make it a tailored tool to your discipline.
>>17023095>k-means clustering is not a learning algorithmPsychotic delusion.See:https://www.mit.edu/~9.54/fall14/slides/Class13.pdfhttps://en.wikipedia.org/wiki/K-means_clusteringhttps://en.wikipedia.org/wiki/Unsupervised_learninghttps://www.ibm.com/think/topics/k-means-clusteringOr basically any other source that covers this subject.
>>17023096From your own source>The unsupervised k-means algorithm has a loose relationship to the k-nearest neighbor classifier, a popular supervised machine learning technique for classification that is often confused with k-means due to the name.Imagine being this retarded lmao.
>>17023097I like how your mental retardation is so profound you think the sentence does something other than to spell out your mistake.
>>17023099you know you're talking to a bot, right?
>>17023099>>17023101Wrong.
>>17023101Maybe, but how is it any different to talking to any of the other biological nonsentients on this board?
>>17023103it isn't. maybe just don't reply to total retards?
>>17021937This trilemma is derived from an absolute, divine perspective—one that is non-human and thus beyond our capacity to consider, as it transcends human limitations.Human beings always rely on presuppositions derived from the relationship between their psyche and the culture in which they live—a process in which childhood plays a pivotal role. In other words, we do not merely form conceptual interpretations of sensory impressions and intellectual processes; rather, we do so through the lens of cultural experience.
