nyrikki

66 ms·

811.
by nyrikki 3y ago
Bad title, this was a hybrid ML/human effort, not a ML only achievement. From the article: "What’s most exciting to me is modelling new modes of human–machine collaboration,” Ellenberg adds. “I don’t look to use these as a replace
812.
by nyrikki 3y ago
Lots of words have odd spelling due to false etymologies. Scissors, Island, Ache, and Could; were all from old English with random letters added by people in the 15th century who mistakenly thought they were from Latin. Etymology is hard, a
813.
by nyrikki 3y ago
Random cite describing the difference: "Prescriptivism is the term used for approaches to language that set out rules for what is regarded as “good” or “correct” usage. Descriptivism is an evidence-based approach to language that descr
814.
by nyrikki 3y ago
Like "i" before "e", except after "c"? While "i" does come before "e" nearly 75% of the time—even after "c" it is a probabilistic rule and not particularly useful. So not so well d
815.
by nyrikki 3y ago
As someone who suffers from dysgraphia, which isn't well understood and commonly missed or completely misunderstood, our teaching methods simply do not translate to effective methods. In English especially where we have problems with v
816.
by nyrikki 3y ago
A particular subset of Connectivism have a philosophical belief that the mind IS a neutral net, not that it is a reductive practical model. Hinton is one of these individuals and with no definition of what intelligence is it is an understan
817.
by nyrikki 3y ago
Sorry, the way it was described I didn't take as the power set but the constructable sets.
818.
by nyrikki 3y ago
The definable numbers, no matter if you use the constructible, algebraic, or computable numbers are countable infinities, or aleph naught. Any segment of the real line, no matter how small you make it is aleph_1 or an uncountable infinity.
819.
by nyrikki 3y ago
The constructable numbers, which is what you are describing are aleph naught, or countable. The Cantor set is a way to think about this.
820.
by nyrikki 3y ago
Do the same with three or more attractors or exit basins and you will find that even with a 0 or negative Laponov exponent (non-chaotic), some systems are indeterminate. While typically taught as chaotic, Newton's fractal is probably
821.
by nyrikki 3y ago
The real numbers are irrational almost everywhere, they are uncomputable 'almost everywhere' also. “what type of discrete systems give rise to the observed macro behavior” is what we have been doing forever through Laplacian deter
822.
by nyrikki 3y ago
It is chalk and cheese. ML requires training, significant amounts of compute power, and large datasets. The Apollo program used Kalman filters with limited compute resources. Kalman filters are for predicting system states in the presence o
823.
by nyrikki 3y ago
Pretty light on details, but if it works for you that is great. Quartiles and hand counting would be my argument for 12 but am probably biased based on familiarity.
824.
by nyrikki 3y ago
Wasn't the US housing crisis of the late 2000s caused by that 99% threshold? Not in finance at all but I do use reverse Kalman filters, to which this seems similar in core concepts. While reverse Kalman filters are incredibly helpful i
825.
by nyrikki 3y ago
Base 12 being 3-smooth, any number with 2 or 3 as a factor has a reciprocal with a terminating expansion. base 60 is 5-smooth, so any number with 2,3,5 as a factor has a terminating expansion. Just as SI has to add in deg min sec for fields
826.
by nyrikki 3y ago
Entscheidungsproblem, the Halting problem, the total function problem, etc... typically relate to arbitrary programs. The point being is there is no single general algorithm that can solve them. The example you gave above is propositional l
827.
by nyrikki 3y ago
Note this quote from the article. "One way to resolve the paradox, therefore, is to put these types into a hierarchy, so they can only contain elements of a “lower level” than themselves. Then a type can’t contain itself, which avoids
828.
LLMs can't self-correct in reasoning tasks, DeepMind study finds(bdtechtalks.com)
166 points by nyrikki 3y ago|358 comments
829.
by nyrikki 3y ago
Pre-print https://arxiv.org/abs/2310.01798
830.
by nyrikki 3y ago
The halting problem is part of asymptotic analysis on this case. Just like with elementary limits you never reach it but you approach it. The halting problem on what we typically call computers decidable but not in practical time lines. The
831.
by nyrikki 3y ago
Well as all you have as a response is an ad hominem you aren't the intended audience anyway. I would love to learn why I am wrong.
832.
by nyrikki 3y ago
Attention is probably most easily conceptualized as run time reweighting. A powerful tool but it doesn't really change the underlying model it is just modifying the weights at runtime.
833.
by nyrikki 3y ago
The halting problem simply doesn't apply to biological neurons, they don't side step them, unless your context is humans writing down an Algorithm. The rules of a human writing down an algorithm is the same thing as a Turing machi
834.
by nyrikki 3y ago
Everything from ion channel response up is non-linear and not a deterministic Turing machine. Random cite https://www.jneurosci.org/content/13/1/334?ijkey=9f8730dd3d0...
835.
by nyrikki 3y ago
Pairwise linear in respect to inputs, for feed forward networks like those that use attention is the restriction related to linear or linearizable. Those words are important. But we know that ML cannot model all turning machines in the limi
836.
by nyrikki 3y ago
Not quite that simple, or well known. We know that the class of total Turing computable functions is not learnable in the limit. But as to what else is AI-complete is mostly open questions. But there is a difference. Obviously ML is far be
837.
by nyrikki 3y ago
Parrots, in the context of stochastic parrots is in reference to repeating or imitation mechanically without understanding. The 'common sense' problem in AI is the important part, not the as of yet undefined abstract concept of in
838.
by nyrikki 3y ago
No, this paper doesn't change that claim at all. This paper is about learnability and not generalization of what is learned. From the papers summary, note that they are referencing learning and not some implications that it gets us clo
839.
by nyrikki 3y ago
The lack of generally accepted definition for knowledge and intelligence makes these conversations difficult but the above post is important to remember. The paper has nothing to do with the common sense problem. While it is a preprint, I d
840.
by nyrikki 3y ago
Feed forward networks are pairwise linear in respect to inputs, they are effectively DAGs. In theory you could represent an entire LLM as a single 2 dimensional graph of linear line segments. It wouldn't be useful for much as the param

More ›