2024
principal component analysis in space forms is accepted for publication as a regular paper in the ieee transactions on signal processing
i gave a talk on universal representation of permutation-invariant functions on vectors and tensors in the 35th international conference on algorithmic learning theory (alt), san Diego
i presented a poster on universal representation of permutation-invariant functions on vectors and tensors in information theory and applications (ita) workshop, san Diego
our paper an effective neural model for circuit netlist representation got accepted in the 27th international conference artificial intelligence and statistics (aistats)
our paper learning ultrametric trees for optimal transport regression got accepted in the 38th annual aaai conference on artificial intelligence
our paper universal representation of permutation-invariant functions on vectors and tensors got accepted in the 35th international conference on algorithmic learning theory (alt)
our paper optimal tree metric matching enables phylogenomic branch length reconciliation got accepted in research in computational molecular biology conference (recomb)
i will teach dsc 40b, theoretical foundations of data science ii, in the upcoming winter quarter
2023
i presented a poster on sum-decomposable models for identifiable set functions in tilos retreat and industry day, la jolla, san diego
provably accurate and scalable linear classifiers in hyperbolic spaces is accepted for publication in knowledge and information systems
principal component analysis in space forms is available on arxiv
2022
i presented posters on principal component analysis in hyperbolic spaces
tilos retreat and industry day, la jolla, san diego
encore fall retreat, la jolla, san diego
i will teach dsc 10, principles of data science, in the upcoming fall quarter
learning hyperbolic embedding for phylogenetic tree placement and updates is accepted for publication in biology
hyperaid: denoising in hyperbolic spaces for tree-fitting and hierarchical clustering is accepted for publication in the proceedings of the international conference on knowledge discovery & data mining (sigkdd)
halıcıoğlu data science institute (hdsi) fellowship award to work with profs. yusu wang and siavash mirarab, ucsd
phylogenetic placement problem: a hyperbolic embedding approach is accepted for publication in the 19th annual satellite conference of recomb on comparative genomics (recomb-cg)
i began a postdoctoral position with prof. siavash mirarab at univeristy of california san diego (ucsd)
2021
on november 2, i successfully defended my doctoral dissertation title: machine learning in space forms: embeddings, classification, and similarity comparisons doctoral committee: ivan dokmanic, olgica milenkovic (advisors), bruce hajek, and maxim raginsky
highly scalable and provably accurate classification in poincare balls is accepted for publication in 21st ieee international conference on data mining (icdm) [arxiv]
my talk on hyperbolic distance geometry problems [abstract, presentation] mini-symposium on sensor network localization and dynamical distance geometry the fields institute for research in mathematical sciences
on procrustes analysis in hyperbolic space is accepted for publication in the ieee signal processing letters
2020
oral and poster presentation of hyperbolic distance matrices at kdd 2020 (virtual event)
kdd 2020 registration award