syadegari

After reading the Tensorflow white-paper (http://download.tensorflow.org/paper/whitepaper2015.pdf), I realized that I lack understanding of the many topics discussed in the paper. Examples of this include Cross-Device and Multi-Device Communication, Graph Partitioning, Partial Executions and Device Constraints and Optimization (most of section 5 subsections). I want to invest in resources, like courses/books or projects, that help me understand these concepts better. My background is in numerical methods and optimization. I am comfortable with topics discussed in an undergraduate algorithm course. I also can read/write Python/C for numerical calculations.

I am interested in research topics at the intersection of topology/geometry and computer science. My previous line of research has been focused on numerical methods with application to mechanical engineering problems (Using methods such as Finite Element and Finite Difference). After working for two and half years in industry and being exposed to new computational problems (often with their roots in computer science), I’ve decided to pursue a master degree that combines my interests in geometry/topology and computer science. Some of the key/important topics of interests that I’ve come across so far have been:

1- A short article with three examples of computational homology https://www.math.upenn.edu/~ghrist/preprints/nieuwarchief.pdf

2- https://www.youtube.com/watch?v=8XJes6XFjxM

3- Computable Topology, the study of “topological and algebraic structure of computation”.

4- Topological data analysis (https://www.youtube.com/watch?v=x3Hl85OBuc0), application of geometrical methods in learning algorithms and in natural language processing/parsing (for the last point though I was not able to find a lot of resources).

I appreciate comments or suggestions regarding the following questions:

1- Research Ideas at the intersection of math (geometry and topology) and computer science. This could be a new topic that I haven’t mentioned here or a refinement/variation of the above ideas.

2- Graduate schools with similar research interests, particularly in the Netherlands and the western part of Germany.

3- Given the interdisciplinary nature of the topics I’ve mentioned, would you recommend a master study in a math department or a computer science/informatics department?

As for me, I have obtained my PhD in Computational Mechanics and I am familiar with some topics in applied math (self study during and after my graduate studies) such as Algebra, Linear Algebra, Analysis, Functional Analysis, Variational Methods and Point-Set Topology.