Data Science and Scientific Computing track core course
Instructors: Bernd Bickel, Chris Wojtan, Gasper Tkacik
Teaching Assistant: Ruslan Guseinov, Rok Grah, Ewa Gajda-Zagorska
Description
Format: The course is divided into three 4-week cycles in which students work in interdisciplinary groups under the supervision of a DSSC faculty member. During each cycle, students first learn the necessary background and tools, and are then coached by faculty to tackle a specific DSSC problem or data set in pairs.
Topics:
- cycle 1: numerical simulation of physical systems (B. Bickel)
- cycle 2: understanding and visualizing data (G. Tkacik)
- cycle 3: numerical algorithms (C. Wojtan)
Goals:
- Provide hands-on experience and scientific insight into different DSSC problems and methodologies
- Learn about evaluation criteria for good models in different fields
- Build a community of computational / data students by project work
- Practice the following skills: handling data, extracting knowledge from data, creating models, running numerical simulations, identifying and understanding sources of error, working in mixed background teams, written and oral communication
Requirements/Exams
Evaluation is based on homeworks and written or oral reports at the end of each cycle.
Credits
ECTS credits 6
Final Grade
- homework/exercises (50%)
- project presentations/reports (50%)
Schedule (subject to change)
| Date | Topic | Location | Other |
|---|---|---|---|
| Mon, 26-Feb-2018 (10:15 - 11:30) | Introduction to Scientific Computing, Floating Point Arithmetic and common pitfalls | Mondi 2 | |
| Wed, 28-Feb-2018 (10:15 - 11:30) | Introduction to FEM, Direct Stiffness Method | Mondi 2 | |
| Mon, 05-Mar-2018 (10:15 - 11:30) | Sparsity, Linear Solvers | Mondi 2 | |
| Wed, 07-Mar-2018 (10:15 - 11:30) | ODEs, Time Integration | Mondi 2 | |
| Fri, 09-Mar-2018 (13:00 - 17:00) | Solving Homework 2 together, TAs will be present (attendance is optional) | Lab Building West, meeting room 2nd floor | |
| Mon, 12-Mar-2018 (10:15 - 11:30) | Variational Formulation of Elements | Mondi 2 | |
| Wed, 14-Mar-2018 (10:15 - 11:30) | Stiffness Warping, Randomized Optimization | Mondi 2 | |
| Fri, 09-Mar-2018 (13:00 - 17:00) | Solving Homework 3 together, TAs will be present (attendance is optional) | Lab Building West, meeting room 2nd floor | |
| Mon, 19-Mar-2018 (10:15 - 11:30) | Particle Swarm Optimization, Topology Optimization | Mondi 2 | |
| Wed, 21-Mar-2018 (10:15 - 11:30) | Data Visualization (Guest Lecture by Thomas Auzinger) | Mondi 2 | |
| Mon, 26-Mar-2018 (10:15 - 11:30) | No lecture, time for working on the project | ||
| Wed, 28-Mar-2018 (10:15 - 11:30) | Presentation of intermediate project results | Mondi 2 | |
| Easter Break | |||
| Mon, 16-Apr-2018 (10:15 - 11:30) | Introduction. Building models. Multi-electrode array data. | Mondi 2 | |
| Wed, 18-Apr-2818 (10:15 - 11:30) | Units. Descriptive statistics. Histograms and PDF estimation. | Mondi 2 | |
| Wed, 18-Apr-2018 (11:30 - 12:15) | Project Presentation of Segment 1. | Mondi 2 | |
| Mon, 23-Apr-2018 (10:15 - 11:30) | Estimating error bars. | Mondi 2 | |
| Wed, 25-Apr-2018 (10:15 - 11:30) | Comparing distributions. Correlations. | Mondi 2 | |
| Mon, 30-Apr-2018 (10:15 - 11:30) | Quantifying correlations, autocorrelation function. | Mondi 2 | |
| Wed, 02-May-2018 (10:15 - 11:30) | Covariance matrix, Principal Component Analysis. | Mondi 2 | |
| Mon, 07-May-2018 (10:15 - 11:30) | Basic spectral estimation, Fourier transform. | Mondi 2 | |
| Wed, 09-May-2018 (10:15 - 11:30) | K-means clustering. | Mondi 2 | |
| Mon, 14-May-2018 (10:15 - 11:30) | Independent component analysis intro. | Mondi 2 | |
| Wed, 16-May-2018 (10:15 - 11:30) | No lecture, projects Q & A | Mondi 2 | |
| Wed, 23-May-2018 (10:15 - 11:30) | Introduction to iterated systems | Mondi 2 | |
| Wed, 23-May-2018 (11:30 - 12:30) | Project presentations for Segment 2 | Mondi 2 | |
| Mon, 28-May-2018 (10:15 - 11:30) | Bifurcations | Mondi 2 | |
| Wed, 30-May-2018 (10:15 - 11:30) | Project proposal presentations for Segment 3 | Mondi 2 | |
| Mon, 04-Jun-2018 (10:15 - 11:30) | Conditioning | Mondi 3 | |
| Wed, 06-Jun-2018 (10:15 - 11:30) | Clustering | Mondi 2 | |
| Mon, 11-Jun-2018 (10:15 - 11:30) | Topological data analysis | Mondi 2 | |
| Wed, 13-Jun-2018 (10:15 - 11:30) | Mid-term presentations for Segment 3 | Mondi 2 |
Homework
| Homework Description | Due Date |
|---|---|
| Homework 1/1. Floating point arithmetic, computation of mean and standard deviation. Detailed instructions can be found in the course slides. | March 05 |
| Homework 1/2. The code skeleton is online now. To run it, you need to install the following packages with pip. 1) "pip install -U wxPython" and 2) "pip install PyOpenGL PyOpenGL_accelerate" | March 12 |
| Homework 1/3. Dynamic simulations. Time integration. Detailed instructions can be found in the homework description (pdf) below. | March 19 |
| Project Segment 1. Design a Catapult. Detailed instructions can be found in the .pdf document. | Presentation of intermediate results: March 28. Final presentation: April 18. |
| Homework 2/1. Solve Homework Problems 1 & 2. | April 25. |
| Homework 2/2. Solve Homework Problems 3 & 5 (not 4). | May 9. |
| Homework 2/3. Solve Homework Problems 6, 7 & 9 (not 8). | As agreed with the TA (Rok). |
Segment 1
- Segment1 Lecture1 DSSC.pdf
- Segment1 Lecture2 DSSC.pdf
- Segment1 Lecture3 DSSC.pdf
- Segment1 Lecture4 DSSC.pdf
- DSSC HW2 Instructions.pdf
- DSSC HW2 Unsolved.zipUpdated March 7, 10:00
- Segment1 Lecture5 DSSC.pdf
- Segment1 Lecture6 DSSC.pdf
- DSSC HW3 v2.zipUpdated March 16, 13:10
- DSSC HW3 Instructions.pdf
- Segment1 Lecture7 DSSC.pdf
- Segment1 DataVisualization.pdf
- dssc 2018s task4.pdf
- DSSC Catapult.zip
Segment 2
Microelectrode array data files: trace1.zip trace2.zip
Natural images dataset: ruderman.zip
Some background on Fast Fourier Transform, convolutions etc, from Numerical Recipes in C: here, here, here, and here.
Complete lecture notes with homework. Final Project is included at the end of the notes.
Segment 3
Additional Downloads
To take a look at the additional Downloads, please click here. (you must be logged in!)