University of Twente Student Theses
Optimising node2vec in Dynamic Graphs Through Local Retraining
Goulis, M.A. (2024) Optimising node2vec in Dynamic Graphs Through Local Retraining.
|
PDF
2MB |
| Abstract: | In network representation learning, deep learning techniques have been on the rise recently due to advancements in neural networks. One of the earliest techniques is node2vec, which generates vector representations of nodes in a graph by simulating random walks to capture network topology and node similarities. However, the time complexity of these techniques is high and scales with the size of the networks. This is particularly important in the case of dynamic networks, which evolve over time and appear in a variety of real-world scenarios. There is still no general method to optimize node2vec for dynamic graphs. In this thesis, we introduce a new technique that makes representation learning computationally feasible in large dynamic graphs. We approximate node2vec on dynamic graphs by focusing on the retraining of local areas affected by graph updates instead of retraining on each graph iteration. We show that this approximating implementation of node2vec results in minimal loss of accuracy while achieving significant speedup, making it a strong optimization technique for use in dynamic graphs. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 31 mathematics, 54 computer science |
| Programme: | Applied Mathematics MSc (60348) |
| Link to this item: | https://purl.utwente.nl/essays/103078 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page