CS/Math 4414: Issues in Scientific Computing - Spring 2022
Tue/Thu 5.00 - 6.15 MCB 136 (zoom when necessary, link on CANVAS)
"The purpose of computing is insight, not numbers", R.
Instructor: Alexey Onufriev
Hours: Wed. 2-4 pm, by request.
GTAs: Niteya Shah (Hws and Unix Issues only) and Dan Folescu (Project related issues, including AMBER and visualization)
Office: ask GTA
Niteya' Email: email@example.com
Office Hours: Tue/Thu 3.30 - 4.30 until further notice.
Dan's Email: firstname.lastname@example.org
Office Hours: by request.
- Numerical Mathematics and Computing, Cheney & Kincaid (the basics of numerical methods, a refresher course )
- Computational Physics: Problem Solving with Computers, Rubin H. Landau, Manuel J. Páez and Cristian C. Bordeianu (an intro text with lots of examples worked out in detail. Scientific computing focused on the problem, rather than method )
- Numerical Recipes in C (or C++), W. H. Press, et al. (The bible
of scientific computing for those who approach the subject from the application perspective, i.e. physicists, chemists, etc. )
- The beginner's guide to Mathematica, J. Glynn et al. (A really good, albeit somewhat dated, intro into Mathematica. Lots of cool, easy to follow and non-trivial examples ).
- Introduction to UNIX, D. Schwartz . (A nice thin book with lots of detailed examples. Perfect for those not on first name basis with UNIX )
Class project introduced
The Energy Function to be minimized.
Finding minimum of a function in 1D: Newton's method.
Intro to mutidimensional optimization: steepest descent, conjugate gradient. Notes on why global optimization is so difficult.
Advanced Optimization: Simulated Annealing (lecture notes on Canvas).
Advanced Optimization: Nelder-Mead method (lecture notes on Canvas).
Advanced Optimization: Genetic Algorithm (lecture notes on Canvas).
The new way of supercomputing: GPU. Introduction.
Software Libraries: Num. Recipies, GLS, NetLib.
In class: quick estimates of what is computationally feasible
Computational thinking. Laplace's Daemon. Philosophical notes on computation.
Student presentations of select research papers (see options below).
Student Presentations. One presentation per group, ~25 mins. Each group picks one topic from the list below.
25 points. Due: In class, April 26, carry over to May 3rd.
[Base your discussion on this blog, which gives a 30 thousand feet view of the algorithm and its main chievements ]
Discuss "Folding at home" (what that is, the main idea, success stories, current state).
Boskovic et al. ``Protein Folding Optimization using Differential Evolution Extended with Local Search and Component Reinitialization" [Also, look for other papers that may have cited this one, did anyone move beyond what's in this paper? ]
Quantum Annealing [Start off with a very brief intro into quantum computing, the discuss QA. High level. ]
Ten Simple Rules for Making Good Oral Presentations
Evaluations of student talks
UNIX Basics. Each student submits 3 separate solutions via Canvas.
Homeworks. Group work (unless otherwise stated). **Each** group member submits groups' work (PDF), with his/her name, amd names of all the partners indicated. Individual work (when specified): each student submits his/her work (PDF). Submission to canvas in all cases.
HW1. Explore a model of the pandemic. Due Feb 14, 11 pm. GROUP WORK. Submit via Canvas
HW2. Playing with the energy (objective) function: simple examples. INDIVIDUAL WORK. Due March 28, 11 pm. Submit via Canvas.
HW3. Pitfalls of numerical optimization. INDIVIDUAL WORK. Due TBA, 11 pm. Submit via Canvas.
Homework Solution Hints will be available on CANVAS within about a week after due date.
Brief project info. Read this first.
What to present, when, and how
Final report template
Final report template files
Project Milestones and Due Dates:
Project-specific notes and deliverables.
- Brainstorming results due in class, Due Feb 17. No more than 15
min. per group presentation + 3 min discussion
with the rest of the class. For details,
including the format, and how many points you can earn,
see "what to present, when and how", above.
For deliverables, see "Starting notes" below.
- Project Stage I (Brainstorming) Comments. To be revealed after all groups have presented.
- Mid-term progress reports (15 mins. per group. + 2 mins for questions)
due in class, March 31. For details of
the format, and how many points you can earn,
see "what to present, when and how", above. Deliverables are outlined in
"Starting notes", and detailed in Mid-term specific notes below.
- Project Stage II Evaluations. To be revealed after all groups have presented.
- AMBER tutorial. Due within one week after the in-class help session on it. DEADLINE: April 12, 11 pm. This is a pass/fail part of the progress report, needed to get full points on it. Group work, but each student submits one printed report via Canvas
- Pre-final progress reports due in class, April 28 For details,
including the format and how many points you can earn,
see "what to present, when and how", below.
- Final report: see "what to present, when and how", above. Deliverables are outlined in "Starting notes", and detailed in Final report
specific notes below. Submit via
VT Canvas. Due by May 9, 9pm. No exceptions under any circumstance.