I am a postdoc member at the Institute for Advanced Study at Princeton and Perimeter Institute in Canada. I got my Ph.D. in astrophysics in 2021 at Johns Hopkins University. I love physics, math, philosophy, classical music, kendo, and anime.
I am interested in using statistical analysis to understand our Universe. My interests range from observational cosmology to stellar physics and planets. My work led to the discovery of special stars that shine out of gravitational sedimentation and the cosmological applications of a new statistic that borrows ideas from convolutional neural networks.
Extracting information from stochastic fields is a ubiquitous task in science. However, from cosmology to biology, it tends to be done either through a power spectrum analysis, which is often too limited, or the use of convolutional neural networks, which requires large training sets and lack interpretability. I showed that a new powerful tool called the scattering transform stands nicely between the two extremes. I studied in depth how to intuitively understand this new statistic which is unfamiliar to most physicists, and used cosmology as an example where its performance is on a par with neural networks while being well-structured statistics. My paper won an award of astrostatistics and was reported by this astrobites blog.
To advocate the use of the scattering transform, I wrote a publicly available module
ST.py based on pytorch, which can implement 1D and 2D scattering transform in a fast and transparent way.
I also work on white dwarfs, the destiny of most stars in the universe. Using data from the Gaia space mission, I discovered a new population of white dwarfs that cool extremely slowly and some others that are merger products. My work has led to two papers (click the figures below), one of which was highlighted by astrobites and AAS Nova and considered a major discovery in Gaia data.
To make white dwarf research easier, I also built a publically available package
WD_models in python for transformation between white dwarf photometry and physical properties.
In high school, my twin brother (who is studying philosophy now) and I found an efficient way to take spectra of meteors with digital camera. We designed a prism device that can screw in front of a lens. We ordered several from a factory, and sold them to other amateurs of astronomy. Shown below is one spectrum of the Geminid meteor shower, taken in 2010. We have made a new batch of such prism devices with a cost of less than 200 dollars each. If you are interested, please contact us or buy it here!
School of Natural Science, Astrophysics Group
2022-, visiting fellow
Department of Astronomy
Advisor: Prof. Eric Peng (彭逸西)
Cheng, S., Two delays in white dwarf evolution revealed by Gaia, 2019, Proceedings of IAU, 15(S357), 175
Liu, D. Z., et al., Potential scientific synergies in weak lensing studies between the CSST and Euclid space probes, 2022, arXiv:2210.16341, accepted for A&A
Lu, C. X., Schlaufman, K. C., Cheng, S., An Increase in Small-planet Occurrence with Metallicity for Late-type Dwarf Stars in the Kepler Field and Its Implications for Planet Formation, 2020, AJ, 160, 253
+1 443 207 1532
1 Einstein Dr, Institute for Advanced Study
Princeton, NJ08540, USA