jmmcd

15 ms·

121.
by jmmcd 5y ago
> Suppose you’re hiring someone for a job that requires a few different mental tasks. (Arithmetic, sequential memory, whatever.) If you knew someone’s g, you could guess how well they’d do at each task. But it would only be a guess! To r
122.
by jmmcd 5y ago
I thought the exact opposite. The narrator thinks the gen-z site is ugly, busy, whatever, and then realised that's what a segment of the market wants. You left out the "and then" part of the story.
123.
by jmmcd 5y ago
Would it be better to have something like a plain text page to hold the outline, with links to individual files?
124.
by jmmcd 5y ago
I've seen online Word become extremely slow even for small documents with a few editors. It doesn't help that there are lots of versions of online Word - there's the one that lives in Sharepoint, the one I get to via Outlook,
125.
by jmmcd 5y ago
I suppose the question is where you want the sorted version to appear. In the output of `ls`? You could write a wrapper to `ls`, maybe?
126.
by jmmcd 5y ago
There sure is a lot of squinting going on in this thread.
127.
by jmmcd 5y ago
Yes it is! I calculated R = 0.58, 0.81, 0.92, respectively. These will be highly significant, comparable to the p-value in the headline. https://docs.google.com/spreadsheets/d/1QTEB7KF1PXBV_TjgBsuV...
128.
by jmmcd 5y ago
Research isn't mentioned, but this is good advice for research too. If you want to make a contribution to research you have to know the field, and be able to re-implement what others have done, before you can add anything.
129.
by jmmcd 5y ago
I guess the author's response would be no, it would still give us cars, but they would have good wheels instead of square ones.
130.
by jmmcd 5y ago
Yes - to be clear, the author knows this is the confounder. That's why it was silly of the author to claim that the paper would describe a correlation without a trivial confounder.
131.
by jmmcd 5y ago
This is rather silly (and not in the way intended). The goal of the paper is to demonstrate a strong linear correlation which arises without a trivial confounder (such as age influencing both reading age and shoe size in children). However,
132.
by jmmcd 5y ago
That article doesn't mention the normal distribution. I suppose she referenced it because she hoped people would assume that an article on an important topic wouldn't just be thrown around by the same author as a citation in suppo
133.
by jmmcd 5y ago
She's not proposing a falsiable hypothesis here which can compete in the marketplace of ideas. We're talking about the normal distribution . If she doesn't know what it is, that casts doubt on her understanding of all the ot
134.
by jmmcd 5y ago
> she’s clearly qualified to participate in a scientific discussion No, her remarks about the normal distribution disqualified her from all scientific discussion.
135.
by jmmcd 5y ago
Agreed... but the decision trees we learn from data to do classification are quite different from the decision trees we design to formalise our processes. I think several replies here have missed the distinction.
136.
by jmmcd 5y ago
The bigger and more interesting point is how to navigate the world of research where everyone is convinced their way is the only right way. I paragraph-TOCs, but my reviewer insists on including it. This article is an example. We need a bet
137.
by jmmcd 5y ago
I don't want to review for a conference with no page limits.
138.
by jmmcd 5y ago
This is not "meaningful to women", it is meaningful to all.
139.
by jmmcd 5y ago
Nobody is saying the most important thing about a politician is their sex. It's not suggested by any of the commentary here either, except the trolls like yourself.
140.
by jmmcd 5y ago
It's a large and gpu-hungry model.
141.
by jmmcd 5y ago
Do you really think Humphries just decided to introduce a bill while McEntee is on leave? Surely this bill would've been in preparation for months.
142.
by jmmcd 5y ago
I disagree. If we consider a single customer with a particular application, we probably imagine an iterative process like this: the customer supplies some sample images, the system gets a few wrong, they add special cases, fine-tune the hyp
143.
by jmmcd 5y ago
Nobody ever taught me step 2, and I learned step 1 from the IT Crowd (ok, not really). Seriously though: are there are any other steps that everyone knows?
144.
by jmmcd 5y ago
The idea of the kangaroo being blind and "feeling around" might be a good metaphor for an optimiser which uses numerical estimation of the gradient, but backprop does not do that. At a stretch, one could argue that being blind is
145.
by jmmcd 5y ago
If we were trying to describe stochastic gradient descent , this would be relevant, but we're talking about backprop (which does often use just a batch, but that's not inherent). And there is nothing about backprop that makes the
146.
by jmmcd 5y ago
Key quote which shows that this young person was a problem-solver, even if his computer skills weren't elite: > I created a “good” BASIC file from scratch [...] Then I compared it to one of the “bad” ones.
147.
by jmmcd 5y ago
> In standard backprop, the most common NN training method, the kangaroo is blind and has to feel around on the ground to make a guess about which way is up. A major problem with standard backprop is that the distance the kangaroo hops i
148.
by jmmcd 5y ago
Ah, yes, that's what I meant. Thanks.
149.
by jmmcd 5y ago
No, I mean the text that was already italic in the original.
150.
by jmmcd 5y ago
I love CM, but the italic Lorem Ipsum sample on this page is uuuugggggly, so bad that it must be a bug. I hope. I'm on MacOS, Firefox 88. EDIT no wait, I see what has happened. It has turned the emphasised text into double-italics, ins

More ›