I will be joining Walmart Global Tech’s search ranking team as Senior Data Scientist.
I earn my PhD in computer science at Rensselaer Polytechnic Institute, advised by Prof. Alex Gittens. From summer 2020 to summer 2023, I was an scholar of the RPI-IBM Artificial Intelligence Research Collaboration(AIRC) program, under the supervision of Shashanka Ubaru and Lior Horesh.
Research Interest
My research focuses on using randomized linear algebra(RandNLA) tools to quantifiably trade off accuracy with computation time. My work lies in obtaining low-rank approximations for ML models(e.g. Matrix completion, Tensor Decomposition). I am also curious about exploring the possibilities of leveraging these RandNLA techniques in Deep Learning settings.
Here is my CV.
Publications and Preprints
Active Learning for Kernel Ridge Regression(Under Review)
Dong Hu, Alex Gittens, Malik Magdon-Ismail
Association for the Advancement of Artificial Intelligence (AAAI 2025)
Provable fast and convergent Low Rank Tucker Decomposition via sketching(In preparation)
Alex Gittens, Dong Hu, Shashanka Ubaru, Vassilis Kalantzis, Lior Horesh
Transactions on Machine Learning Research (TMLR 2024)
Sparse graph based sketching for fast numerical linear algebra[arXiv]
Dong Hu, Shashanka Ubaru, Alex Gittens, Kenneth L. Clarkson, Lior Horesh, Vassilis Kalantzis International Conference on Acoustics, Speech and Signal Processing (ICASSP 2021)
NoisyCUR: An algorithm for two-cost budgeted matrix completion[arXiv]
Dong Hu, Alex Gittens, Malik Magdon-Ismail
European Conference on Machine Learning (ECML-PKDD 2020)
Misc. interests
I love traveling and taking photos. Most of my recent photography work is done using either medium format or large format film cameras. You can find most of my works (in low resolution) on my instagram account. You can always reach out to me to acquire the raw files for archival-quality prints.
I also engage in Texas Hold’em poker during my free time, playing both online and offline. I am a profitable poker player at GGpoker(NL200 Zoom level).