Thema der Dissertation:
Machine Learning for Cancer Survival Prediction Thema der Disputation:
Learning on Graphs: Node Representation Learning with Node2vec
Machine Learning for Cancer Survival Prediction Thema der Disputation:
Learning on Graphs: Node Representation Learning with Node2vec
Abstract: Graphs are powerful data structures that can capture interactions between objects, such as relationships between individuals in a social network or interactions between proteins in a protein-protein interaction network. However, these interactions make machine learning on graphs a challenging task because objects cannot be regarded as independent entities as in most standard machine learning tasks. Instead, node representation learning methods can be used to learn low-dimensional vector representations that capture interactions between objects, which are represented by nodes, and their structural context within the graph.
In my first talk, I will discuss node representation learning using node2vec as a prime example. Node2vec employs biased random walks in conjunction with the skip-gram model to generate node embeddings that can effectively capture different types of network neighborhoods of the nodes in the graph. After explaining the methodology of node2vec, I will also briefly discuss its advantages and limitations.
The second talk will be a summary of my dissertation, in which I explored machine learning for cancer survival prediction. In this talk, I will first introduce my research questions and then briefly discuss the methodology and findings of my dissertation with respect to each of these questions.
In my first talk, I will discuss node representation learning using node2vec as a prime example. Node2vec employs biased random walks in conjunction with the skip-gram model to generate node embeddings that can effectively capture different types of network neighborhoods of the nodes in the graph. After explaining the methodology of node2vec, I will also briefly discuss its advantages and limitations.
The second talk will be a summary of my dissertation, in which I explored machine learning for cancer survival prediction. In this talk, I will first introduce my research questions and then briefly discuss the methodology and findings of my dissertation with respect to each of these questions.
Zeit & Ort
05.07.2024 | 13:00
Seminarraum 049
(Fachbereich Mathematik und Informatik, Takustr. 9, 14195 Berlin)