snikolov

20 ms·

121.
by snikolov 16y ago
I agree -- artificial intelligence modeled on human brains hasn't been explored nearly enough to label it a failure. In addition to advances in artificial neural networks, there has been a lot of progress in more realistic models of the br
122.
by snikolov 16y ago
I think you're right. The brain is a very specialized computing machine, and it evolved for a specific purpose.* Evolution doesn't reward general computing capabilities. Hence, computers are vastly more general than brains. If something
123.
by snikolov 16y ago
Disclaimer: I haven't read the book (but am now planning to) and may be completely off mark with my comments, but here's what I make of all this. Sure, there are certain things that are not computable. This is a theoretical limitation on
124.
by snikolov 16y ago
I agree. I also don't think one can get very far in that respect without first understanding the computational principles and mechanisms by which the brain can perform ordinary tasks that are extremely difficult for computers (e.g. vision)
125.
by snikolov 16y ago
I believe Minsky is arguing that because these "devices of calculation" are not well understood, you cannot conclude that there is anything special about them that puts them beyond the reach of any computational (algorithmic) processes. I t
126.
by snikolov 16y ago
Another interesting side note is that wavelets (Gabor wavelets in particular) have been shown to look very much like receptive fields in the primary visual cortex. The suggestion is that such receptive fields evolved as an efficient way to
127.
by snikolov 16y ago
The network effects of going to a good school are enormous. I don't mean just who you know, but also what they can teach you (mostly outside of class). At MIT, I've found that classes have mostly taught me specific course material, while
128.
by snikolov 16y ago
That's a very good point. This seems to be a major point against an explanation based solely on encouragement/bias. In a solely encouragement/discouragement based theory, it might be that for some reason, males are encouraged as well as d
129.
by snikolov 16y ago
I have a feeling that the variability is at least partially modulated by factors such as (lack of) encouragement and gender bias. Imagine a spinning flywheel (this is a terrible analogy but I think it will illustrate my point) with particl
130.
by snikolov 16y ago
Eliminating 6.001 also seems to me a part of a trend to allow more freedom in the EECS curriculum. Aside from eliminating 6.001 (which is the most talked about change), computer science students now have more freedom to take computer scien
131.
by snikolov 16y ago
I think characterizing the courses that replaced 6.001 as exercises in "pasting libraries together" is unfair. The new introductory courses 6.01 and 6.02, are focused on introducing fundamental EECS design principles, in the context of mob
132.
by snikolov 16y ago
Steven Strogatz has studied (and to a large extent popularized) mathematical models of synchronization of fireflies (and many other things). Here's an interesting TED talk by him about synchronization in general: http://www.ted.com/talks/
133.
by snikolov 16y ago
I think getting your hands dirty with something you are interested in is the key here. Too many classes in "software engineering" deal with contrived examples built around concepts you are "supposed to be learning." More freeform projec
134.
by snikolov 16y ago
A little perspective from a junior in EECS at MIT who is also concerned about leaving college as a competent software engineer. In most software jobs, practical experience outweighs theoretical knowledge. I would guess that the people from
135.
by snikolov 16y ago
Check out http://www.autonlab.org/tutorials/ and also a free ebook called Elements of Statistical Learning http://www-stat.stanford.edu/~tibs/ElemStatLearn/ I've also found a number of course note sets helpful, for example, MIT's machin
136.
by snikolov 16y ago
If you ask someone knowledgeable a specific question, they can frame the answer (and question) in the right context, customized for you and your question, and make you see the big picture. Also, you can follow up with more questions until
137.
by snikolov 16y ago
Speaking of high school grade systems, my school had grades out of 100. Talk about unnecessary competitiveness (both with yourself and others). I obsessed over getting 97s and 98s rather than 94s and 95s. While we weren't officially rank
138.
by snikolov 16y ago
It's kind of sad that it goes from 'struggling to comprehend superstring theory' to being an unknown speck at <random company>.
139.
by snikolov 16y ago
more time working does not necessarily amount to more productivity.
140.
by snikolov 17y ago
While there's somewhat of a culture of being "hardcore" and taking lots of classes at MIT, I find that many people abandon that mentality after their freshman year. In reality, most people are just trying to get by with the minimum 4 class
141.
by snikolov 17y ago
Signals and Systems problem set http://mit.edu/6.003/S10/www/handouts/hw3.pdf