brosco

5 ms·

1.
by brosco 11mo ago
One reason is that it would be like hanging a picture using a sledgehammer. If you're just studying various ways of unwrapping a sphere, the (very deep) theory of manifolds is not necessary. I'm not a cartographer but I would assu
2.
by brosco 11mo ago
Good point! I used to be guilty of this myself, so now I'm pretty sensitive about other people doing it. I am now one of the more senior students in an academic research group, and some of the younger members would benefit from this ad
3.
by brosco 11mo ago
I'm not saying it's a learning method. And I don't see how anyone could mistake this for science, so why would it be pseudoscience? It's not really about effort either. It's just a trick that helps me pay attention
4.
by brosco 11mo ago
I have a tip for following lectures (or any technical talk, really) that I've been meaning to write about for a while. As you follow along with the speaker, try to predict what they will say next. These can be either local or global pr
5.
by brosco 11mo ago
One of OpenAI's founding team members developed Adam [0] well before it was flashy and profitable. It's not like nobody is out there trying to develop new algorithms. The reality is that there are some great, mature solvers out th
6.
by brosco 1y ago
There is indeed a lot of crossover, and a lot of neural networks can be written in a state space form. The optimal control problem should be equivalent to training the weights, as you mention. However, from what I have seen, this isn't
7.
by brosco 1y ago
Very cool. Analysis I was the first "real" math textbook that I (an engineer, not a mathematician) felt like I could completely follow and work through, after a few attempts to get through others like Rudin. Hopefully the Lean com
8.
by brosco 1y ago
Have you heard of JIT libraries like numba ( https://github.com/numba/numba )? It doesn't work for all python code, but can be helpful for the type of function you gave as an example. There's no need to rewrite
9.
by brosco 1y ago
Compared to Matlab (and to some extent Julia), my complaints about numpy are summed up in these two paragraphs: > Some functions have axes arguments. Some have different versions with different names. Some have Conventions. Some have Con
10.
by brosco 1y ago
This looks great! I've been needing something like this for a while, for a project which is quite compute-heavy and uses lots of threads and recursion. I've been using valgrind to profile small test examples, but that's defin
11.
by brosco 2y ago
That's a good observation, and it is indeed true for many Markov chains. But your counterexample of the identity matrix is not quite right; every vector is an eigenvector of the identity, so there is no "realignment" needed.
12.
by brosco 3y ago
It's absolutely being explored. There is a lot of active research into using ML to learn solutions of PDEs (Navier-Stokes in this case). It's not my field so I don't know much about the specifics. The works that I've rea
13.
by brosco 3y ago
If you're interested to learn more about aerodynamics I would highly suggest learning a bit of classical aerodynamics. It will not be software oriented, since most of the theory deals with approximating very complicated behavior with s
14.
by brosco 3y ago
Thanks for clarifying the motivation, that makes a lot of sense.
15.
by brosco 3y ago
I guess this will probably come up in the reviews but the presentation of the Kalman filter is lacking. I know it's not the point of the paper, but getting these details wrong in a paper about Kalman filters is not encouraging. The sta
16.
by brosco 3y ago
What? If you start light like Wendler recommends, the program is completely manageable. In fact, most people I know think there is too little training volume at first. I used it for several months in a row a few years ago and it led to grea
17.
by brosco 3y ago
I've had my Framework 12 over a year now, so maybe I can give some perspective. I had a lot of similar issues (mostly with wifi) running Fedora when I first got it. I also felt like I wasted a lot of time getting it set up and fixing l
18.
by brosco 3y ago
More specifically, using mixed integer linear programming. I've never seen an MILP used this way, to characterize the entire feasible set (or "solution pool"). Is this one of the fastest ways to do so? The usual branch-and-bo
19.
by brosco 5y ago
My favorite of the year was Anna Karenina. If you like Brothers Karamazov it would be right up your alley.
20.
by brosco 5y ago
In addition to Boyd and Vandenberghe, I like "Lectures on Modern Convex Optimization" by Ben-Tal and Nemirovski. Particularly the section comparing linear and conic optimization problems.
21.
by brosco 6y ago
When applied to a linear system with Gaussian noise, the particle filter will (approximately) recover the solution of the Kalman filter. So yes, they solve different problems, but the particle filter is more of a generalization of the Kalma