RESEARCH
My doctoral research explored how deep representation learning and submodular selection can recommend relevant, complementary references for an academic manuscript.
If Google Scholar is temporarily unavailable, you can read the publications below and open their original sources.
RESEARCH FOCUS
I study how language representations and selection algorithms can help people navigate large collections of information. My citation-recommendation work brings relevance, coverage and diversity into the same decision.
PHD THESIS · 2022
PhD thesis at the University of Technology Sydney. The work combines learned document representations with set selection, considering the usefulness of a reference list as a whole rather than relevance rankings alone.
Read thesis at UTS (opens in a new tab)SELECTED PUBLICATIONS
Explore the linked sources for full papers and publication details.
2021
IEEE Access
A jointly trained neural representation and submodular scoring model for selecting citations that balance relevance and diversity. Evaluated on three citation datasets.
2020
ICPR 2020 · author preprint
Uses Sentence-BERT with Siamese and triplet training to represent papers, then selects a reference set with a submodular objective. Evaluated on the ACL Anthology Network corpus.
2020
PACLING 2019 · revised selected papers
A reference-selection method that considers relevance, coverage and diversity together, with experiments using an ACL publication corpus.
Let’s talk about information retrieval, recommendation, applied AI or a research collaboration.
