dvse

11 ms·

31.
by dvse 15y ago
Not sure this is good advice - without background I have mentioned there is no chance of actually making sense of the material in the CMU course - note how linear algebra is a prerequisite and they cover convex optimization at the very begi
32.
by dvse 15y ago
Start with the MIT linear algebra course (18.06) by Gilbert Strang and Stanford course on linear dynamical systems (EE263) by Stephen Boyd. Then move on to Boyd's course on convex optimization (EE364). Lectures for all of these are on youtu
33.
by dvse 15y ago
With a fair bit of hand waving it's also possible to present the law of large numbers and the central limit theorem in the same way. Can also look at Metropolis-Hastings as modifying a suitable random walk to get the steady state that we wa
34.
by dvse 15y ago
Couldn't you do something like this with Kinect? Depth information is all one needs to get a plausible "software focus" effect. Not sure how effective it would be outdoors though.
35.
by dvse 15y ago
This is one of the first observations someone learning the subject would make.
36.
by dvse 15y ago
The only thing this graphic does is that it generates confusion and reinforces misconceptions. A tree diagram is completely inadequate for this task - they say that mathematics is about "analogies between analogies" for a reason.
37.
by dvse 15y ago
As a quick rule of thumb would be to never trust anyone who claims "regression" is a separate topic from "classification". Oh yes, and in this case it is beyond awful.
38.
by dvse 15y ago
Notice that to get a "step" plot it is necessary to assume that probability of getting an A is strongly correlated with the SAT score. The paper says that for humanities there is no strong correlation. On the other hand, the "result" for ph
39.
by dvse 15y ago
They don't show it. The obvious explanation is that the grades in humanities are not as strongly correlated with SAT.
40.
by dvse 15y ago
Why is a physics professor surprised by this? You will get exactly the same "step" plot directly from the binomial distribution if you assume independent trials and a linear dependency between the SAT band and the probability of getting a
41.
by dvse 15y ago
One way to demonstrate your knowledge in a technical field like engineering or computer science would be to publish some research papers in well known journals or conferences. By no means perfect but much better than nothing. This is not
42.
by dvse 15y ago
There might be more to selling ads than initially meets the eye - the ability to connect specialized products and services with interested audience can help drive all sorts of innovation quite unrelated to social media. Many of these projec
43.
by dvse 15y ago
For what it's worth, statistical machine learning and other CS people have recently "discovered" classical convex analysis - for the last few years papers have been full of stochastic subgradients and the like. The next big thing has got to
44.
by dvse 15y ago
Thanks for all your contributions! Indeed, a rare perspective on HN (or anywhere for that matter).
45.
by dvse 15y ago
You have it - either obliteration or (as an optimistic scenario) total control enacted with the justification of preventing splinter groups from achieving said obliteration. And while someone should think about these things I do admit that
46.
by dvse 15y ago
Did Himmler have the technology to continuously monitor movements and verbal and written communications of every citizen?
47.
by dvse 15y ago
One of the problems with CS PhD graduates at least is the lack of broad familiarity with their and related fields. Outside of a handful of top US schools the resources (i.e. high level graduate courses taught by people who actually understa
48.
by dvse 15y ago
Say what you will, but every ingredient for the world RMS has described in "The Right to Read"[1] is by now well in place. "Pragmatic" line of thinking will likely land us in the most restrictive police state in history. 1. http://www.gnu.
49.
by dvse 15y ago
Certainly nothing remotely comparable to SICP exists. From my point of view people who are interested in these topics are much better off getting some intuition for basic problems in scientific computing and then making the requisite conn
50.
Video lectures on advanced robotics on OCW(ocw.mit.edu)
1 points by dvse 15y ago|0 comments
51.
by dvse 15y ago
I'm not sure it is wise to post links to the site on open forums.
52.
by dvse 15y ago
I never understood the popularity of their AI text - the discussion of topics other than the most basic search methods is uniformly obtuse and the authors hardly ever make any of the important connections with literature outside their "fiel
53.
by dvse 16y ago
This kind of latent variable model is not only highly unstable but also tends to tell you more about the person interpreting the results than the data itself, e.g.: Chang et al., Reading Tea Leaves: How Humans Interpret Topic Models, http
54.
by dvse 16y ago
Regarding "pseudoteaching" mentioned in the article - every great teacher is to some degree a con artist. They know that in the end it is up to the student to engage with the material - the only thing they can do as a teacher is to communic
55.
by dvse 16y ago
What's remarkable about (non-retrospective) "best paper" awards is how poorly they stand up to time. It would be very interesting to add citation counts to get a crude indication of how good people are at identifying significant (or even ju