My research aims to study relevant large-scale social phenomena using computational methods grounded in machine learning, natural language, network science and social psychology. My past and present work includes a study of over 100,000 real-life online discussions to uncover popular topics of contemporary moral concern, how moral dilemmas are framed, and how they are perceived and judged by internet users.
Prior to joining Penn I was an M.Phil. student working with Lexing Xie and Colin Klein at the Computational Media Lab, the Australian National University. I was also affiliated with the Humanising Machine Intelligence Grand Challenge at ANU.
Prior to that I pursued B.S. in computer science at the School of Computing and Information Systems, the University of Melbourne. During my undergraduate studies, I worked with Charl Ras at the School of Mathematics and Statistics on designing resilient network embeddings.
I also have a special interest in optimization, especially in the context of machine learning. I have collaborated with Nguyen Tran at the School of Computer Science, the University of Sydney, on designing personalized and communication-efficient algorithms for federated learning.
I publish under my official Vietnamese name, Tuan Dung Nguyen.
[2023-09-12] We release a new language model for astronomical research called AstroLLaMA. Joint work with Yuan-Sen Ting and Jo Ciuca at the ANU Research School of Astronomy and many others at UniverseTBD. A short technical report can be found here.
[2022-03-16] Our paper, Mapping Topics in 100,000 Real-life Moral Dilemmas, is accepted to ICWSM 2022! The code can be found here and data here.
[2021-11-30] I am honored to attend the International School in Artificial Intelligence and its Applications in Computer Science (ISAAC 2021) this december at Monash University.
[2021-04-27] I will be volunteering at the ICWSM-2021 conference.
[2021-04-16] I am honored to be selected to the Cornell, Maryland and Max Planck Pre-doctoral Research School (CMMRS-21).
[2020-12-15] Check out our pre-print DONE: Distributed Newton-type Method for Federated Edge Learning on arXiv.
[2020-10-26] Our work, titled Personalized Federated Learning with Moreau Envelopes, is accepted to NeurIPS 2020. The code can be found here.
[2020-05-26] Our paper, Federated Learning with Proximal Stochastic Variance Reduced Gradient Algorithms, is accepted to ICPP 2020.