Elise Sänger's home page

E-mail:    sanger at strw.leidenuniv.nl
LinkedIn:   https://www.linkedin.com/in/elise-sanger/ 
Elise Sänger

About me

I am a master student in astronomy and cosmology at Universiteit Leiden. I am curious about the world around me and want to understand how it works on a fundamental level. I hope to one day be able to make a significant contribution to our understanding of the Universe. I am especially interested in gravitational wave physics and what we can learn about the universe from gravitational waves.

As a person, I am hard-working, a fast learner, and have an analytical and helicopter way of thinking which helps me with getting an overview of a situation without losing sight of the details. Although considered a nerd, I am creative and pragmatic when solving problems which can lead to unexpected solutions. I can work independently and I am also a good team worker. I strive to attain the highest achievable in any situation.

Born in Sweden, living in the Netherlands, and having worked in Switzerland, I love working in an international environment. Getting to know people from all over the world and learning from different cultures helps making this world a better and more peaceful place.

Research

Master thesis

Gravitational Wave Echoes in Black Hole Quasi-Normal Modes
Abstract: With the first detection of gravitational waves in 2015, a whole new way to observe the Universe has opened up. One of the things that has become possible to study is what happens close to the horizon of black holes. This can be done by looking at gravitational waves from the ringdown phase of binary black hole coalescences. For classic black holes, the waveform of the ringdown phase consists of the quasi-normal modes of the final black hole. Various theories for quantum black holes predict structure near the horizon of a black hole, which leads to gravitational wave echoes in the ringdown phase. Abedi, Dykaar and Afshordi reported finding tentative evidence for the presence of gravitational wave echoes. We investigated this result by searching for echoes in the data of GW150914. The analysis presented in this thesis did not find any evidence for the presence of gravitational wave echoes following GW150914. Our analysis indicated that the noise levels were too high to be able to detect the presence of echoes with an amplitude similar to the best fit value that Abedi, Dykaar and Afshordi found.
Gravitational wave echoes

First master research project

The Alignment Between Brightest Cluster Galaxies and the Large-Scale Structure in the Hydrangea Simulations
Abstract: Observations show that the orientations of brightest cluster galaxies (BCGs) are roughly aligned with the orientations of their host clusters. To understand the origin of this alignment the alignment angles between the BCGs and their clusters for the 24 massive clusters in the Hydrangea simulations was analysed. The alignment between BCGs and their clusters is already present at high redshifts and becomes slightly tighter towards present time, when the average alignment angle is 11°. This implies that the orientations of BCGs and clusters can be explained by a combination of primordial alignment and tightening of the alignment over time by tidal fields.
Brightest cluster galaxy

Bachelor thesis

Graph Challenge: Counting Triangles and Enumerating Trusses in Graphs
Abstract: Counting triangles and enumerating trusses are fundamental graph analysis algorithms with several applications. The 2018 Static Graph Challenge sought to improve the efficiency of these algorithms, so that the time for graph analysis decreases and larger graphs can be analysed. In this thesis we compared different serial algorithms on the CPU for both triangle counting and truss enumeration. The algorithms for triangle counting can be divided into two categories: (1) linear algebra based algorithms using the adjacency matrix of a graph, and (2) graph based algorithms using the edge list and neighbourhoods of a graph. We concluded that the graph based algorithms were faster and could handle larger graphs than the linear algebra based algorithms. For enumerating trusses we concluded that treating 3-trusses as a separate case gave better results because special properties of 3-trusses could be used. We also concluded that the runtime for k-trusses did not strongly depend on k for k > 3.

After writing our own algorithms we looked at the champions of the 2018 Graph Challenge, as well as the results for triangle counting of the 2017 Graph Challenge. The focus of the champions was on improving the rate of edges processed per second by improving how the graph data is partitioned for parallel computing and how the graph data is stored in memory. One of the champions did this for linear algebra based algorithms on CPUs by improving the partitioning of the adjacency matrix before doing masked sparse matrix matrix multiplication. Another champion did it for a graph based algorithm on GPUs by improving how the graph partitions are stored in memory before and during processing. The highest rate of edges processed per second achieved for triangle counting in the 2018 Graph Challenge was 4.7 · 109 edges per second, which is approximately a factor 104 higher than the rate that we achieved for the same graph. The largest graph processed for triangle counting in the 2017 Graph Challenge has 2.7 · 1011 edges, which is approximately a factor 105 larger than our largest graph. These numbers are not comparable due to differences in available computing power.
Trusses