Some thoughts wrt my classes. First of all, I have spent way too much time thinking about what classes to take. There are lots of cool classes, the optimal set of classes to take is hard to tell. And furthermore, classes are cool and important and useful. But the most important thing is to think about cool problems.
Anyways, I think I’ve arrived at a pretty optimal location.
chemistry:
- I will go to lecture and read the lecture notes during lecture. and also pay attention in class unless there is a good reason not to. But also not worth stressing too much here. But ofc I should do good in the class. It is a task that I should do.
- participating in class will make me enjoy it more
chinese + philosophy
- they are fun classes and help me think about stuff
- they are excuses to spend time developing answers to non-math questions of life.
- participating in class will make me enjoy it more
cryptography:
- to get the most out of it I should read the textbook and participate in class!
demaine:
- lectures: just watch when you have some free time
- problems: work on as much as you can!
- in class: work with cool ppl (don’t actually discriminate too
- much here, probably most of the ppl in the class are cool) on
- try to find some problem that I especially care about and work on it more. develop into a research-y thing
graph theory: grind every day with friends and by myself. as long as I stay on top of it it’ll be great. don’t stress about star problems.
alloc: keep pushing on this, its a great problem
keep collecting and creating cool problems!
misc:
- good to spend time with ppl sometimes, especially if its
- special it is more treasured than if every day can become too
- mundane
don’t waste time.
don’t wait ’till older or “opportunity” to do cool research! just do it right now!
more specifics
Here is what I’ve learned so far
crypto:
- computational intractibility lets us do stuff
- PRGs are cool if they exist
- you can do length extension, and that’s pretty good
- next-bit unpredictibility is same as “pseudo-random”: kinda crazy
algo sampling: - pretty cool idea: “coupling markov chains” to make them converge
fun with hardness: - 3-partition is strongly hard - some variants of this: - assume exactly 3 in each set - 3DCM - 3X - partition is weakly hard
ham cycle is hard
SAT: lots of variants
- Not-All-Equal SAT
- 1-in-3 SAT
lots of variants are in P rip - 2 SAT - horn SAT
gtac: - projective plane incidence graph is a cool graph - the graph is a bipartite graph on e.g., lines and points contained in the lines
- looking at why inequalities are loose is a good idea
- as always prob-method OP