About me

I am an Assistant Research Scientist in Physics & Astronomy at Johns Hopkins University, working with Prof. Alex Szalay and Prof. Rosemary Wyse. I use machine learning to address fundamental questions in astrophysics — from the chemical composition and evolution history of the Milky Way and M31, to the physics of the first billion years of the universe.

As a member of the JHU Galactic Archaeology group, I develop spectral foundation models (SpecCLIP and LoRA fine-tuning, and the successor) to align and transfer stellar information across surveys (LAMOST, Gaia XP, DESI, APOGEE, PFS, etc.), estimating chemical abundance for Galactic archaeology/Near-field cosmology. I am also a formal member of the Subaru Prime Focus Spectrograph (PFS).

During my PhD at Tsinghua University (advisor: Prof. Yi Mao), I studied the Epoch of Reionization — the cosmic dawn when the first stars ionized the intergalactic medium. I built tools including simulation-based inference, wavelet scattering transforms, and diffusion models to extract cosmological information from faint 21 cm data. I also spent 18 months as a visiting researcher at Institut d'Astrophysique de Paris and developed explainable machine learning techniques. (mentor: Prof. Benjamin D. Wandelt).

Experience

  • Nov 2025 –
    Assistant Research Scientist, Johns Hopkins University
    With Prof. Alex Szalay & Prof. Rosemary Wyse
  • Aug 2024 – Nov 2025
    Postdoctoral Fellow, Johns Hopkins University
    PFS science return · Mentors: Szalay & Wyse
  • Jul – Aug 2024
    Visiting Researcher, NAOC / UCAS, Beijing
    Host: Prof. Yang Huang · Stellar spectra foundation models
  • Nov 2022 – May 2024
    Long-term Visiting Researcher, Institut d'Astrophysique de Paris
    Mentor: Prof. Benjamin D. Wandelt · Explainable ML for astrophysics

Education

  • 2018 – 2024
    PhD in Astronomy, Tsinghua University
    Thesis: Exploring the Epoch of Reionization with Machine Learning · Advisor: Prof. Yi Mao
  • 2014 – 2018
    B.S. in Physics, Wuhan University

Publications

Refereed Journals
Generalization from Low- to Moderate-Resolution Spectra with Neural Networks for Stellar Parameter Estimation: A Case Study with DESI
X. Zhao, Y.-S. Ting, R. F. G. Wyse, A. S. Szalay, Y. Huang, L. Dobos, T. Budavári, V. Wei
ApJ 1006, 2 (2026)
SpecCLIP: Aligning and Translating Spectroscopic Measurements for Stars
X. Zhao, Y. Huang, G. Xue, X. Kong, J. Liu, X. Tang, T. C. Beers, Y.-S. Ting, A.-L. Luo
ApJ 998, 189 (2026)
Ordered Embeddings and Intrinsic Dimensionalities with Information-Ordered Bottlenecks
M. Ho, X. Zhao, B. D. Wandelt
Machine Learning: Science and Technology (2025)
Simulation-based Inference of Reionization Parameters from 3D Tomographic 21 cm Light-cone Images. II. Application of Solid Harmonic Wavelet Scattering Transform
X. Zhao, Y. Mao, S. Zuo, B. D. Wandelt
ApJ 973, 41 (2024)
Square Kilometre Array Science Data Challenge 3a: Foreground Removal for an EoR Experiment
A. Bonaldi, …, X. Zhao, …
MNRAS 543, 2 (2025)
Can Diffusion Model Conditionally Generate Astrophysical Images?
X. Zhao, Y.-S. Ting, K. Diao, Y. Mao
MNRAS 526, 1699 (2023)
Implicit Likelihood Inference of Reionization Parameters from the 21 cm Power Spectrum
X. Zhao, Y. Mao, B. D. Wandelt
ApJ 933, 236 (2022)
Simulation-Based Inference of Reionization Parameters from 3D Tomographic 21 cm Lightcone Images
X. Zhao, Y. Mao, C. Cheng, B. D. Wandelt
ApJ 926, 151 (2022)
How to Evaluate the Sufficiency and Complementarity of Summary Statistics for Cosmic Fields: An Information-theoretic Perspective
C. Sui, Y. Mao, X. Zhao, T. Jing, B. D. Wandelt
ApJ 998, 2 (2026)
Conference Proceedings
Finetuning Stellar Spectra Foundation Models with LoRA
X. Zhao, Y.-S. Ting, A. S. Szalay, Y. Huang
ICML ML4Astro Workshop · Spotlight (2025)
3D ScatterNet: Inference from 21 cm Light-cones
X. Zhao, S. Zuo, Y. Mao
ICML ML4Astro Workshop (2023)
Evaluating Summary Statistics with Mutual Information for Cosmological Inference
C. Sui, X. Zhao, T. Jing, Y. Mao
ICML ML4Astro Workshop (2023)
Machine Learning and the SKA for Cosmic Dawn and the Epoch of Reionization
A. Acharya, …, X. Zhao
Advancing Astrophysics with the SKA II (AASKAII)
Under Review & Invited Chapters
Foundation Models for Astrophysics
X. Zhao, Y.-S. Ting
Invited chapter — "ML Techniques for Astrophysics and Cosmology"
Galactic Archaeology with the Subaru 'Ōnohi'ula Prime Focus Spectrograph Strategic Program
M. Chiba, …, X. Zhao
PFS Strategic Program overview
Likelihood-free Model Selection in Cosmic Reionization with 3D Tomographic 21 cm Lightcone Images
T. Binnie, X. Zhao, J. R. Pritchard, Y. Mao
Submitted to AAS Journals (2025)

→ Full record on Google Scholar

Selected Talks

Jul 2026
2nd Annual Cosmic Horizons ConferenceContributed · UVA, USA Decoding the Chemical Fossil Record: Machine Learning and Foundation Models for Near-Field Cosmology
May 2026
LSST Discovery Alliance Regional MeetingContributed · JHU, USA Cross-Modal Learning for Galactic Archaeology: Prospects in the Rubin Era
Apr 2026
BCCP / Cosmology SeminarSeminar · UC Berkeley, USA Decoding the Chemical Fossil Record: Machine Learning and Foundation Models for Near-Field Cosmology
Apr 2026
Center for Decoding the Universe Journal ClubJournal club · Stanford, USA Decoding the Chemical Fossil Record: Machine Learning and Foundation Models for Near-Field Cosmology
Jan 2026
247th AAS MeetingContributed · Phoenix, USA SpecCLIP: Aligning and Translating Spectroscopic Measurements for Stars
Jul 2025
ICML 2025 ML4Astro WorkshopSpotlight · Vancouver, Canada Finetuning Stellar Spectra Foundation Models with LoRA
Apr 2025
JHU/STScI CAS Wine & Cheese SeminarSeminar · JHU, USA From 21 cm Astrophysics to Galactic Archaeology: Enriching Physics-Driven Analysis with Machine Learning
Sep 2023
Astro CoffeeInformal · IAS, Princeton, USA Can Diffusion Model Conditionally Generate Astrophysical Images?

Contact

Department of Physics & Astronomy, Johns Hopkins University · Baltimore, MD 21218