In this talk, we give a quick introduction to embedding nodes and whole graphs in vector spaces [1]. We highlight the connection to word
representation learning and discuss how graph embedding models process the raw graph to gain meaningful representations. We also address two
of our applications. In [2], we collected Ethereum-related data from multiple sources (Twitter, Etherscan, Tornado cash) to deanonymize
Ethereum users. In [3], we collected Twitter data related to Covid-19 and classified tweets based on the expressed vaccine view.
[1] Rozemberczki, B., Kiss, O., & Sarkar, R. (2020). Karate Club: An API Oriented Open-Source Python Framework for Unsupervised Learning on Graphs.
Proceedings of the 29th ACM International Conference on Information & Knowledge Management, 3125–3132.
https://doi.org/10.1145/3340531.3412757
[2] Béres, F., Seres, I. A., Benczúr, A. A., & Quintyne-Collins, M. (2021). Blockchain is Watching You: Profiling and Deanonymizing Ethereum Users.
2021 IEEE International Conference on Decentralized Applications and Infrastructures (DAPPS), 69–78.
https://doi.org/10.1109/DAPPS52256.2021.00013
[3] Béres, F., Csoma, R., Michaletzky, T. V., & Benczúr, A. A. (2021). Vaccine skepticism detection by network embedding.
Book of Abstracts of the 10th International Conference on Complex Networks and Their Applications, 241–243.