jmmcd

16 ms·

151.
by jmmcd 6y ago
> would take way too long I think the right way to do it would be to swipe a few movies every few days, whenever you're wasting time on your phone anyway. That way you'd have a queue of matches ready to go on Friday night.
152.
by jmmcd 6y ago
Hmm, it looks like the discussion about emoji has been removed/moderated from the page? Anyway, my point was only that if someone is unhappy that emojis are being added, then proposing new non-emojis doesn't really solve their pro
153.
by jmmcd 6y ago
> Ambiguity is the norm Right. But as you say, usually our priors are good enough. The cases where we stop, double-take, and deliberately look from another angle are rare.
154.
by jmmcd 6y ago
Off-topic, but the man in the interview reminds me of this lovely song about a foot fetish https://www.youtube.com/watch?v=2XDCoVoeMKs
155.
by jmmcd 6y ago
And there's no reason we couldn't have a deep learning system where the input data (images) included time-stamps and movement vectors, and it could be good both at easy image classification, and at choosing particular "head m
156.
by jmmcd 6y ago
That doesn't help
157.
by jmmcd 6y ago
Really excellent!
158.
by jmmcd 6y ago
I think a doctor would recommend to see a doctor, here.
159.
by jmmcd 6y ago
A line of best fit wouldn't quite make sense to me, as the implication would be that extrapolation on that line would lead to improvements -- forever. A GA allows for iterative generation of multiple faces (no linear assumption) and se
160.
by jmmcd 6y ago
Well, I think they are using a GA, based on: > Machine learning-based algorithms subsequently determined which faces produced the greatest amount of activity for each person, then established which traits those faces had in common. Based
161.
by jmmcd 6y ago
> If the goal is to personalize attractive faces, you could just have subjects rate pictures on a scale of 1-100. However, research in the field (the field here is really "interactive genetic algorithms") suggests that it'
162.
by jmmcd 6y ago
The EEG is just a gimmick, indeed.
163.
by jmmcd 6y ago
Of course all this is USA-specific. In my area of Ireland, around the early 90s, the pattern for all local numbers changed from 091-85xxx to 091-555xxx. You could omit the 091 when dialling local numbers. Since all my friends lived nearby,
164.
by jmmcd 6y ago
Yes, he writes "If you start at x and I start at x then your MCMC run is no better than mine." and surrounding text is making this point.
165.
by jmmcd 6y ago
Yes, I think you're confused about regularisation. A regularisation is (usually) a component in the overall loss, which has the goal of simplifying or preventing overfitting, as opposed to the main component which has the goal of fitti
166.
by jmmcd 6y ago
> It's still staggeringly elitist and unaware to assume that "genuinely" literate people (whatever that means) should not only know of Pratchett The comment you're referring to didn't say that, or suggest it. So
167.
by jmmcd 6y ago
One perspective is: anything that can be automated (thus lowered in cost) should be. For drudge-work, of course that's good. For some examples, showing that it can be automated shows that it IS drudge-work. But replacing a creative ill
168.
by jmmcd 6y ago
I disagree - it's not clear. I thought it was going to be some fancy machine learning to silence keyboard sounds while keeping the mic active for voice.
169.
by jmmcd 6y ago
This is great -- I just used it to draw a (cartoon) objective landscape for a stochastic objective, for my Optimisation class.
170.
by jmmcd 6y ago
I write my lectures and labs in .md and convert to pdf with pandoc. I like the results tex produces but I don't love the language, so pandoc is ideal.
171.
by jmmcd 6y ago
Just arrived in my inbox: IEEE Software CFP for special issue on program repair. The editor is a GP person. https://www.computer.org/digital-library/magazines/so/call-f...
172.
by jmmcd 6y ago
My PhD student is doing GP for (univariate) TS one-class classification.
173.
by jmmcd 6y ago
The terminology is bad. Don't confuse "search and optimisation"-type search with "pathfinding"-type search. A-star is the latter, but GP (and other metaheuristics like GA) are the former. To be clear -- if you'
174.
by jmmcd 6y ago
No, you are missing the thing that sets GP apart from GAs and other metaheuristics. It is the P -- GP is about creating programs.
175.
by jmmcd 6y ago
Counter-point: we all know that "free" internet services are flaky. Eg many people prefer to pay for email, for reliability. This article introduces the perspective that OA journals are a bit like ad-supported email. I don't
176.
by jmmcd 6y ago
We don't always know what is important at the time it's published, and there are many examples of this in the literature. Meanwhile people are working hard to preserve every Commodore 64 game.
177.
by jmmcd 6y ago
Yes, and of course we have arXiv and friends, and sci-hub, and researchers' and institutions' own pages. But all this misses the point a little -- it is not just the articles that should be preserved, but the journal itself, as a
178.
by jmmcd 6y ago
I know what you mean, but no. It wasn't that she had done something wrong, it was just amazing (to her manager) that she had done something so (apparently) brave as reaching out to the FBI. The point of the essay is -- some people are
179.
by jmmcd 6y ago
Well, there's a study in the article where being "ugly" is measured as 10x the effect size of being the "wrong" race. This was in criminal sentencing.
180.
by jmmcd 6y ago
> if you're ugly and rich you're just ugly with money Spending money on the right things can make a large difference. Teeth, skincare, diet, haircut, clothes, time in the gym, quitting a job that would keep you in the sun too l

More ›