geoalchimista

8 ms·

31.
by geoalchimista 6y ago
Second this. I read a few chapters of the book and put it down. I was quite disappointed by its irrelevancy to the current state of the academia. Nothing against his time. It's just that the world has changed dramatically.
32.
by geoalchimista 6y ago
Off-topic: I couldn't help noticing that there is such a recurrent pattern that under every NumPy-related topic, there are fervent disciples of Julia who prophecy that the end (of Python) is nigh and you should all be converting to Jul
33.
by geoalchimista 6y ago
No thanks. Would rather learn Latin, which is a real language.
34.
by geoalchimista 6y ago
> get paid about $50k per year as researchers I quite agree with you that the current system is highly exploitative. But the problem runs deeper within the academia than with the visa system. First, there isn't a market to begin wit
35.
by geoalchimista 6y ago
> The point is that Julia is easier to use and faster than Python for just about any scientific task. Try to make a plot. Last time I checked, the time to first plot was still a pain point of Julia. > it's harder to use and less
36.
by geoalchimista 6y ago
I think you may be overgeneralizing. If the speed of number crunching is the only concern, then sure, few scientists would use Python, R, or MATLAB. This is not the case in reality. Whether you would run into speed limit of course depend
37.
by geoalchimista 6y ago
The title seems to assume that Bayesian statisticians do not know Monte Carlo methods. But in fact, Monte Carlo methods are part of Bayesian statistics. They are in any decent Bayesian stats textbooks.
38.
by geoalchimista 6y ago
Second this. Research code is already hard, and with misaligned incentives from the funding agencies and grad school pipelines, it's an uphill battle. Not to mention that professors with an outdated mindset might discourage graduate st
39.
by geoalchimista 7y ago
Even in a language where writing three layers of loops is an okay choice (e.g., Fortran), I'd argue against using it everywhere because it is poor abstraction and it will likely be less performant than calling the relevant BLAS routine
40.
by geoalchimista 7y ago
Because it took half a year for the paper to get accepted. You can see that the manuscript was sent to Nature Methods on July 28, 2019.
41.
by geoalchimista 7y ago
> Bsc gets paid by EU taxes, guy ends up teaching in US university or working for US company. Why would this be a concern? They should be free to go where they want to. Just because university education is publicly funded does not mean t
42.
by geoalchimista 7y ago
The point I wanted to make was that string "nan" is usually not as useful as floating point NaN when dealing with numerical data. Therefore, as default behavior it is acceptable, if one has to make a choice.
43.
by geoalchimista 7y ago
There are always choices to make. In this case, I would much prefer to let the data be treated as is, i.e., no silent casting of np.nan to the string "nan". When dealing with numerical data, a string "nan" is rarely usef
44.
by geoalchimista 7y ago
I think this blog article did not frame the novelty of this study very well. It is not the nowcast that is the problem, it is getting nowcast at a spatial resolution higher than weather forecast models. But in that case, the limitation is m
45.
by geoalchimista 7y ago
Not the OP, but as a working scientist, it seems to me there are two important issues, among many others. (i) Reproducibility is not a criterion for tenure evaluation, therefore it bears little relevance for career advancement for young sci
46.
by geoalchimista 7y ago
I'm really not so amused by the bloating of words like "machine learning through polynomial regression". People have been doing polynomial regression for centuries (almost as old as linear regression) and how does it become a
47.
by geoalchimista 7y ago
Cultural problem seems to be playing no small part in this. In average high schools, being interested in STEM is just not considered something cool.
48.
Terrence Tao – An Introduction to Measure Theory (2012) [pdf](terrytao.files.wordpress.com)
5 points by geoalchimista 7y ago|0 comments
49.
by geoalchimista 7y ago
A recipe for the typical media hype "AI learns physics" discovery 1. Find a well-known classical problem with a battle-tested deterministic model representation 2. Generate synthetic data using the model 3. Train a 1-layer neural
50.
by geoalchimista 7y ago
In other words, we need an optimizing linear algebra compiler to compile high-level abstractions to efficient BLAS/MKL calls, instead of just interpreting them for a literal translation. For example, in the case of NumPy, when given x
51.
by geoalchimista 7y ago
> In my opinion, the default should always be zero. This won't work well for a lot of cases. Imagine plotting the atmospheric CO2 concentration time series in the past 200 years. Setting the original point at zero ppm would not make
52.
by geoalchimista 7y ago
I think it is well known that floating-point numbers have their limitations. How would the IEEE 754 floating-point system know when to use L'Hôpital's rule? If you know there is a singularity (0 / 0 or inf / inf) then
53.
by geoalchimista 7y ago
Because that would grant the government the power that enables it to become a worse perpetrator even than the big tech?
54.
by geoalchimista 7y ago
Does defining tau := 2 * pi make math easier at all?
55.
by geoalchimista 7y ago
"pro-intellectual" by sending intellectuals to Gulag?
56.
by geoalchimista 7y ago
The link is a preprint. Here is the published version: https://www.tandfonline.com/doi/full/10.1080/07474938.2013.8... .
57.
by geoalchimista 7y ago
Don't know why you are downvoted but I second your opinion. The contents listed here seem to emphasize applications in CS rather than theoretical aspects (e.g., Hilbert space, tensors, spectral theory).
58.
by geoalchimista 7y ago
It's true but photosynthesis has two stages. The light reaction stage is quite efficient. Chlorophyll a is slightly more efficient than commercial solar panels. The dark reaction stage, which assimilates CO2 to produce carbohydrates, i
59.
by geoalchimista 7y ago
"2 electrons with 1 photon" --- that is how photosynthesis works in nature. Great thinking!
60.
by geoalchimista 7y ago
One thing I can think of is that R is not well suited for computationally expensive work --- by "computationally expensive" I mean something that takes months to run on a cluster. Not only is R 100--1000x slower than Fortran, but

More ›