counters

17 ms·

241.
by counters 5y ago
To be extremely pedantic, their general approach isn't new. Folks in the meteorology community have been toying with generative modeling approaches for the past five years. For example [1] used GANs for super-resolution reconstruction
242.
by counters 5y ago
Can you share who your vendor/host is? In another comment you mentioned that you're mmap'ing files on SSDs... on vendors like GCP that sort of hardware (especially the storage volume you'd need to have access to a handfu
243.
by counters 5y ago
Not all of them. But many are. It's important to be skeptical of claims made about th novelty and quality of weather forecasts from new/unknown players.
244.
by counters 5y ago
Water vapor is, indeed, comprehensively included in climate models since the 1970s.
245.
by counters 5y ago
Bingo. And it goes vice-versa; because software engineering standards are so lax (read: non-existent) in most academic disciplines and domains, on the rare occasion where you get accompanying code to an academic work it may be impossible to
246.
by counters 5y ago
The thought could be that in "tree form", the material contributes to the net uptake of carbon from the atmosphere, so once it is harvested that negative flux on the balance sheet shifts equilibrium more towards lower or net negat
247.
by counters 5y ago
It's worth pointing out that there likely is no such thing as the AMO. The scientist who coined the term in some pioneering work in 2000 has, over the past few years, been extremely outspoken that it was probably an errant interpretati
248.
by counters 6y ago
There's been some great, recent efforts to bring more "modern amenities" to the Fortran coding experience. For instance the Fortran Package Manager [1] aims to make it easier to bootstrap new applications leaning on establish
249.
by counters 6y ago
I can second this. Exact same policy for our data science program's recruitment pipeline; a PhD is a hard requirement and Insight was the only program we noted when reviewing applications. Still, it would be a toss-up between the PhD&#
250.
by counters 6y ago
Usually if you use Cython, f2py, numba, or something similar, the goal isn't to create code that fits into the ecosystem of the compiled language - it's to optimize a kernel that fits into the Python ecosystem with minimal baggage
251.
by counters 6y ago
!00% nailed it. ML/AI are tools. As in any other exercise, tools can make it easier for beginners/amateurs to engage with a project, but they don't replace 10,000 hours of experience and deep domain experience and understandi
252.
by counters 6y ago
> Basically 2010-2018 was an open field for ML/DL research with old(ish) methods being rapidly applied to low hanging fruit and large datasets with newly cheap compute. Absolutely! And what many folks need to continue to remember is
253.
by counters 6y ago
But is it a benefit or a liability? Reminds me a lot of the arguments in Susskind and Susskind's "The Future of the Professions."
254.
by counters 6y ago
You hit on a really important point here, the notion of "dread risk"[1]. The public tends to poorly characterize and amortize the different degrees of risk associated with different activities, due to biases in how we perceive the