I'm a master student currently studying in UBC (Vancouver, Canada).
My research topic is implicit representation and its application in astronomy. I'm also interested in applying deep learning techniques in biological studies.
Apply computer vision techniques (e.g. neural fields) on astronomical observations.
Apply deep learning on protein studies (e.g. alphafold)
University of British Columbia | Vancouver, Canada
2022 - Now
University of British Columbia | Vancouver, Canada
2019 - 2022
Xi'an Jiaotong-Liverpool University | Suzhou, China
2015 - 2019
University of British Columbia | Vancouver, Canada
May 2021 - PresentTrained a Multilayer perceptron (MLP) to map celestial coordinates (RA/DEC) to pixel values, i.e. an implicit neural field that encodes galaxy distributions as network weights.
Further incorporated transmission function of imaging sensors and trained the MLP to map from RA/DEC concatenated with wavelength to hyper-spectral cubes.
Calculated pixel values via integrating hyper-spectral cubes over the transmission functions and approximated using Monte Carlo approximation with importance sampling on wavelength.
Performed single-shot inpainting using the MLP and outperformed state-of-the-art models (e.g. Self2Self). Our model can inpaint with decent quality (>40 PSNR) images under bands with no observations at all.
Alida | Vancouver, Canada
Sep 2020 - Apr 2021Worked on an agile team that is responsible for maintaining and developing a software that enables easy creation, distribution, and analysis of surveys.
Contributed to the debugging of customer reported issues. Routinely helped with QA and maintain CI/CD pipelines.
Veritatis et quasi architecto beatae vitae dicta sunt, explicabo.
Sed ut perspiciatis, nam libero tempore, cum soluta nobis est eligendi.
Excepteur sint occaecat cupidatat no proident, quis nostrum exercitationem ullam corporis suscipit.
Enim ad minim veniam, consectetur adipiscing elit, quis nostrud exercitation ullamco laboris nisi.
Maxime placeat, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.
Optio cumque nihil impedit uo minus quod maxime placeat, velit esse cillum.
