j2kun

54 ms·

721.
by j2kun 9y ago
In the context of the post this is the very first "lesson" the author displays. He then expands on his claim with a bunch of metrics. There is nothing to infer hyperbolicity, no implied sarcasm or joke, and any well-intentioned wr
722.
by j2kun 9y ago
What does that even mean
723.
by j2kun 9y ago
Don't know why this is downvoted. If the title is the hardest part of professional blogging, then the content must be relatively vapid.
724.
by j2kun 9y ago
It relies on a quantity being defined which happens to be equal to the mean, but that value can be arrived at without having defined the mean a priori. The minimizers of (7,11) with respect to the 1-norm defined in the post include all valu
725.
by j2kun 9y ago
To this book's credit, it covers a LOT more material than Axler's book. And it's more heavily focused on optimization methods for which the details of tabular representations of matrices takes center stage.
726.
by j2kun 9y ago
What kind of second-rate mathematician tries to prove a theorem without writing down examples? That's Proving Theorems 101.
727.
by j2kun 9y ago
The more I read Nautilus the less I like it. It seemed good initially, and now it feels like bad pop science. Maybe that has something to do with paying writers. I prefer Quanta Magazine.
728.
by j2kun 9y ago
Folks are downvoting this, but they must clearly have drowned themselves in kool-aid. You want fast matrix multiplication? Pretend you're doing texture shading in a language primarily used for dynamic web frontends!
729.
by j2kun 9y ago
In addition to this, there is a known effect of surge pricing that routes demand from poorer areas to wealthier areas. I can't find the paper right now, but IIRC it was 2015.
730.
by j2kun 9y ago
> It is worth noting, by the way, that BitTorrent users were free-loaders, using massively more bandwidth than the vast majority of Comcast’s customers. I don't think that's how a service agreement works. If I'm paying for
731.
by j2kun 9y ago
My question for the author: does implementing a proof formally in Coq honestly give you a better understanding of the theorem/proof? It would seem to me that, if you're filling in details from an existing proof or making some exis
732.
by j2kun 9y ago
> What's really amazing to me is that Stephen Kleene probably proved this without the help of a computer, but I was able to verify it with the highest possible scrutiny. It's as if he wrote an impressive program without ever ha
733.
by j2kun 9y ago
To be fair, your original comment did not mention a disorder, and a reasonable person could assume a lack of a disorder from the comment alone.
734.
by j2kun 9y ago
Recently saw chalktalk[1] which is designed for a different goal (and very rudimentary from what I've tried), but seems to be a step in that direction. [1]: https://github.com/kenperlin/chalktalk
735.
by j2kun 9y ago
It makes me wonder if there are any realms of wizardry where the highest ranking wizards spend all their time on management skills and never get to practice their magic. "Oh, these days I just review a lot of spell-design scrolls and m
736.
by j2kun 9y ago
Unfortunately, it is known that many classes of problems can be learned with adversarial manipulation of an epsilon-fraction of the training data. See for example the paper "PAC learning with nasty noise," though there are many ot
737.
by j2kun 9y ago
Yes, many models are vulnerable. But there are no proofs that allow you to say that. You should also be aware that they're only talking about image classification in all of these. I will only accept blanket statements about "all m
738.
by j2kun 9y ago
Take care, the original authors were being overly broad when they implied "all ML algorithms" can be fooled in this way. Not that I disagree with your conclusion. Just the hypothesis.
739.
by j2kun 9y ago
> No [machine learning] system is 100% secure. That is what's being claimed without proof, when they meant "existing neural network systems are known to be insecure"
740.
by j2kun 9y ago
Well the article asserts this, so I'm asking for clarification on what systems they were thinking of.
741.
by j2kun 9y ago
What "high risk" real world systems use deep learning?
742.
by j2kun 9y ago
> Maybe if you gave the first trees more weight, though That is exactly what boosting does...and one would weight the kernel computation in the same way.
743.
by j2kun 9y ago
Do you have any insights as to why random forests are better (are they better?) in this context (or any context?) than boosted decision trees?
744.
by j2kun 9y ago
Folks here assume "text editing" is just structured code editing. I use vim for more than just editing code trees in one language. I edit unstructured text (lists, notes, articles), semi-structured markup (latex, markdown), and ma
745.
by j2kun 9y ago
Looks like it's failing to process some standard tex commands (e.g. \textup) as well as some user defined macros. See the many display errors in https://www.arxiv-vanity.com/papers/1710.07406/ Of course, it g
746.
by j2kun 9y ago
Another example is James Kilgore, a convicted (second degree) murderer/kidnapper/activist who was on the lam and is now a professor at UIUC: https://en.wikipedia.org/wiki/James_Kilgore
747.
by j2kun 9y ago
There is little analytical or "detailed" about this post. The most complex model is y = 3*x, the author provides no evidence to back up any claims about adoption, difficulty of use, etc., and most of the author's complaints b
748.
by j2kun 9y ago
I think these are not comparable. Tensorflow is a software library, Azure is a compute environment which allows one to run, among many other libraries, tensorflow implementations of ML models.
749.
by j2kun 9y ago
Who makes the decision on what gets standardized into the distro?
750.
by j2kun 9y ago
It's at least a little bit ironic that on the same day major news outlets were announcing the death of a celebrity before he had actually died, and the story stayed up for hours.

More ›