I’m a math student in a pre PhD math program. I love proofs and for a long time I was positive I wanted to get my PhD in pure math, but at some point in my life I think I would like to work with the math behind neural networks. I’m interesting in computer decision making and whether or not computers can learn from their mistakes. While TCS seems to be more focused on...? Well that I’m not sure.
Perhaps learn machine learning subreddit would be a better place to go.
I used the incremental voronoi algorithm, I call these computational Pepes. I consider Pepe to be about frogs, not something controversial. I'm quite proud of these. Should I make more, or show these somewhere? I want to sell them.
MUCHO $$$$$$$ ????
Hi everybody, I'm a CS student and I'm highly interested in develop an academic life. I love algorithms, data structures and theoretical CS. In this year I have to select a topic for make my graduation proyect, but I'm blocked because I don't know what to choose. Recently I've been interested in computational geometry, but none of my professors works in this, so I would like to receive an opinion from somebody who work in this field. Is it too mathematical and have a low amount of algorithms? How hard is to grow academically in that topic? Is hard make papers on it? I ask this because I'm a little afraid that computational geometry turns to be a topic more apropiate for mathematicians. Also I would like to know if some of you have experience in the complexity theory too(same questions). I know that machine learning can be more friendly with CS people, but I'm not too interested on that
I am currently a software engineer writing your standard CRUD apps, but would like to make a career transition into the more "computational geometry" space. To me this would be something like route/path planning jobs: Software Engineer (C++), Routing & Remote Assistance, working on software such as geometric kernels, isometric solvers, etc.: Full Stack CAD web Developer, working on software that does geometric processing: Software Engineer - Geometry, Geometric Software Engineer.
I'm pretty early in the process of this. So my question is, how do I break in to this field? Are the variety of jobs I listed too broad for a single person to be competitive for at once? How do I get started learning what I need to become competitive.
Thanks everyone!
EDIT: I have a masters in CS, I've been working as a software engineer for about 1 1/2 years.
Hello, I've started this book, but as usual, there are no solutions provided, and there's no separate solutions manual available either.
https://www.goodreads.com/book/show/316275.Computational_Geometry
I was wondering if we could work together on producing these solutions?
Maybe a separate subreddit could be setup for solutions to exercises in books, where we can post specific problems and verify each other's solutions for correctness.
Let me know what do you think
My teacher went over in class how to use the method of sweeping lines to solve the following problems but I couldn't follow it at all.
Basically he would plot the conic section; draw a tangent line at point (x,y) where the integer solution clearly exists; then he would "sweep the line" across the section and find the rest of the integer solutions somehow...
Here are some example problems:
- Prove that there are infinitely many positive integer triples (x,y,z) such that x^2 +2y^2 = 3z^2
(we could divide z^2 on both sides to get a conic section in terms of u and v; thus two variables and we can use method of sweeping lines)
- Prove that there are infinitely many primitive (prime numbers only) pythagorean triples.
Hi!
I'm a current undergraduate applying to PhD programs. I was wondering if anyone knows if doing a PhD in a topic like computational geometry can lead to a career in a research lab like the Pixar research group. Or do they want people who are working on a more specific topic within computer graphics?
I was learning the Rotating Calipers technique and I came across his paper here.
On page 87 of the file (p. 79 of the paper), an algorithm to find all antipodal pairs is presented. Someone copied it verbatim and posted on Wikipedia. Here he wrote:
> ANGLE(m,n) is a procedure that returns the clockwise angle swept out by a ray as it rotates from a position parallel to the directed segment P_m P_{m+1} to a position parallel to the directed segment P_n P_{n+1}
Using this definition, I don't think the algorithm will work at all even from the first while loop.
I = 1
J = 2
WHILE(ANGLE(I, J) < pi) DO J := J+1
The correct definition, I believe, should be "counterclockwise angle". Consider ANGLE(1, 2), that's the angle swept from vector P1 P2 to vector P2 P3. If this angle was swept clockwise, definitely it'd be larger than pi, so basically, the while loop would fail at the first check. It should be counterclockwise. Am I right?
Even if ANGLE is actually "counterclockwise", the next steps of the algorithm are also very "weird". What is the intuition behind the variable "CURRENT"? Is it actually necessary or complicating a simple idea? I tried to code based on his method, but I could not produce a correct result.
The idea of Rotating Calipers is absolutely correct and very elegant. I have coded the algorithm in my own style using his ideas and it yields correct results. It's just that I found his particular implementation wrong. I doubt my own judgment since it's a 40-year-old paper, I must have missed some detail. Can you help me verify? Thank you in advance!
P.S.: I also have further concerns about this pseudocode, but the definition of ANGLE bothers me the most.
Hello guys,
I am considering taking a Computational Geometry course at my University. Do you guys think this topic would be helpful for doing CV research or application?
Here is the summary of the course
Algorithms and data structures that are used to solve geometrical problems. Topics include geometric searching, convex polygons and hulls, Voronoi diagrams, plane sweep algorithms, proximity and intersections. Application areas which are discussed include computer graphics, VLSI design, and graph theory.
Thanks for any suggestions and inputs!
ML has already dominated the domain of images. Would Computational Geometry, its algorithms and theorems too would become obsolete? If so how soon? If not, why?
I am debating between the two classes. I know Automata is friday and like over 2 hours long and Computational is T/Th and fits in with the rest of my classes on T/Th where I wont have to drive 25 minutes to campus on Friday if I take it. I also heard Automata is very boring. I haven't heard anything about Computational Geometry but Rosen's rate my professor seems good and the material online doesn't seem too bad.
I'm working on a passion project at work. I'm in manufacturing/engineering and cs is not my major. I took C in college and am working within python now. Or trying to atleast. I have a list of xy coordinates that relate to the machine the cutting path to take. It is cutting out of a thin circular disk many different small pieces. My question is this, is there any way that you've seen or can think of that s program can identify the points at which a piece is cut free? Like once you've cut on all sides it has no Support and falls free. Sometimes the path surrounds the piece in order and sometimes the path intersects with previously cut area to release the piece. I'm just looking to bounce ideas as I'm stuck and my brain is fried from thinking about this from too many wrong angles. Also python and SymPy which is what I'm using may not be the proper tools. I'm unsure. Thank you for your time.
