usgroup

75 ms·

541.
by usgroup 4y ago
As a hiring manager making a fairly low volume of hires annually I always told the candidates why. I also did this in all F2F interviews. If the candidate didn’t make it, Id tell them and then take them through what we thought and why they
542.
by usgroup 4y ago
> What happen when your model ends up in production ? It will start affecting the behaviour of the very thing you're trying to predict Indeed, again part of the reasons I think MLOps is a problem. One ideally needs a causal theory a
543.
by usgroup 4y ago
> MLOps is about enabling reproducibility of previous outputs; it has nothing to do with reinforcing any pattern. This sentence seems to contradict itself. Otherwise, I'm familiar with MLOps :-) Wikipedia describes it as the interse
544.
by usgroup 4y ago
I think that MLops somewhat subdues the modelling and research effort in commercial settings. At least anecdotally I only seem to see the coexistence of ML pipelines paired with the “feature engineering” + fit(X,y) approach to data “science
545.
by usgroup 4y ago
WHO page on breastfeeding: https://www.who.int/health-topics/breastfeeding Too clever by half springs to mind with this article.
546.
by usgroup 4y ago
A sucky team will make everything fail. that a team is like some kind of vat of raw potential waiting for a leader with the right incantation to unlock it, is wonderful, hopeful and — much of the time — wrong.
547.
by usgroup 4y ago
I’d be surprised if it’d be useful for something like cars or soccer players, or really anything that may not have a continuous mapping. I guess more generally whenever the underlying “true” similarity function is not differentiable — categ
548.
by usgroup 4y ago
To answer my own question: https://qdrant.tech/
549.
by usgroup 4y ago
I’m familiar with metric learning within the Mahalanobis family for kNN oriented applications . I’m not getting what use cases this framework targets? Is it custom image search type stuff which may benefit from fine tuning? What is a realis
550.
by usgroup 4y ago
Personally I appreciate the sarcasm —- for me, it speaks to the very often shoddy analysis that ends up in common circulation.
551.
by usgroup 4y ago
"What underlies learning in the brain might be actually simpler than previously thought despite the brain being one of the most complex objects in the known universe. A collaboration of Scientists led by the EPFL Blue Brain Project has
552.
by usgroup 4y ago
It’s a good way to learn Haskell and Prolog —- so many of the problems lend themselves to combinatorial or recursive solutions.
553.
by usgroup 4y ago
I think you’ve got to call the bluff and try to level up. Progress through Physics isn’t monotonic and a bit like the speeding cars of the freeway, you’ll probably meet those going faster at the traffic lights due to the throughout bottlene
554.
by usgroup 4y ago
https://gpytorch.ai/
555.
Owl – Scientific Computing for OCaml(ocaml.xyz)
4 points by usgroup 4y ago|0 comments
556.
Hylang – Lisp for Python v1.0 alpha ready(docs.hylang.org)
20 points by usgroup 4y ago|1 comments
557.
by usgroup 4y ago
Feels appropriate to say that Hylang is in version 1.0 alpha: https://github.com/hylang/hy/releases/tag/1.0a4 That's Lisp over Python -- you can do data science properly in Lisp now using all you fa
558.
Hello and goodbye to the J language (2021)(emiruz.com)
1 points by usgroup 4y ago|0 comments
559.
by usgroup 4y ago
I appreciate the post in the sense that it is an insightful perspective that he didn’t have to share. If you are elite it is difficult to talk about your options, pay or way of thinking without it coming across as nothing but hubris to the
560.
by usgroup 4y ago
This seems like well reasoned advise, I might try it myself! In the intermediate ratings between 1500-2000 what would you say are the biggest factors affecting performance?
561.
by usgroup 4y ago
I don't think that is correct. The analysis explicitly matches pairs of players which have played each other in both Blitz and Rapid time control. Only the Blitz rating difference is used to estimate probability of a win under both tim
562.
by usgroup 4y ago
I found the conclusion surprising personally. Ceteris paribus it isn't obvious that someone who beats you at Blitz wouldn't just beat you even more convincingly given more time. I actually find that a bit mysterious.
563.
by usgroup 4y ago
I think this is an essential part of the analysis. The author makes a point of matching pairs of players who have played each other across multiple time controls so that he is comparing the same population of people across multiple time con
564.
by usgroup 4y ago
> Isn't this pretty much true by definition? Fast thinking is the only thing you can do in quick games, so it has to be more prominent. I think you might be implicitly defining "quick games" as ones where you only have tim
565.
by usgroup 4y ago
It’s an observational study and the author volunteers a bunch of caveats at the end of the article including with regards to construct validity which I think is the crux of your challenge. If you think of chess as two people with different
566.
How much does fast thinking matter in Lichess games?(emiruz.com)
74 points by usgroup 4y ago|45 comments
567.
by usgroup 4y ago
Is anyone using Picat? I'm deciding whether to bother learning it at the moment. I'd be grateful for experience reports.
568.
by usgroup 5y ago
I think Immanuel Kant was around 60 when he wrote the Critique of Pure Reason. There are lots of examples of old people doing good stuff.
569.
by usgroup 5y ago
Thanks I will give this a go. I had a long fling with Lua a while back. Wrote everything in Lua -- it is a great language. I would have used it for data science if only it had more well maintained packages. The main API for Torch was in Lua
570.
by usgroup 5y ago
No, just hypocrites

More ›