Gravitational waves, machine learning, and AI for science

We work at the intersection of gravitational-wave science and AI for Science: how to detect signals, what they tell us about the physics behind them, and what intelligent algorithms can and cannot do in that process.

I am a Research Associate at ICTP-AP, UCAS. The group is small: a handful of students and visitors, working on concrete datasets and questions that can be settled by a test.

Longer notes, talks and papers live on my personal site and blog.

People

Members, visitors and collaborators, newest first.

Boxuan Ge since 2025

Hangzhou Institute for Advanced Study, UCAS · Postdoc

Numerical relativity combined with machine learning.

Xiubo Meng since 2025

Changchun University of Science and Technology · Undergraduate

Diffusion-model denoising algorithms for gravitational-wave data.

Zun Wang since 2025

Beijing Normal University · PhD student (joint training)

Machine-learning waveform template modelling for gravitational waves.

Zhaoren Zhou since 2025

Tianjin University of Science and Technology · Undergraduate

Gravitational-wave detection algorithms based on the fractional Fourier transform.

Xudong Wang since 2025

ITP-CAS · Master's student

Global-fit data analysis and machine learning for space-based gravitational-wave detectors.

Jianan Zhang since 2025

ICTP-AP · Master's student

Galactic-binary foreground search for space-based detectors: global fit and machine-learning methods.

Shuai Pu since 2024

Hangzhou Dianzi University · Undergraduate

Cosmology and machine learning.

Research

Four directions, from gravitational-wave data analysis to the methodology that serves it. Concrete topics include, but are not limited to:

Gravitational-wave science

  • Data analysis and signal detection
  • Multi-messenger data processing
  • Waveform modelling and numerical relativity
  • Detector noise and transient characterisation

AI for science

  • Deep learning for time-series signals
  • Generative models (flow and diffusion models)
  • Physics-informed and operator-learning models (PINN, DeepONet)
  • Simulation-based and Bayesian inference

Intelligent algorithms and automation

  • LLM agents for research workflows
  • AutoML and automated design of scientific models
  • Evolutionary optimisation and heuristic search

Scientific computing

  • Digital signal processing
  • High-performance and GPU computing
  • Scientific software engineering

Emerging directions

  • Quantum computing for scientific computation
  • Complex-system data analysis
Undergraduate research programme

We also run the Taiji Lab undergraduate programme at UCAS, for students who want to work on real data before graduate school.

https://github.com/iphysresearch/UndergradResearchLab/ →

Join us

I welcome students who want to work on gravitational-wave data analysis, machine learning, or AI for science. Strong programmers and self-directed learners are especially welcome.

Backgrounds

  • Physics / astronomy
  • Computer science / software engineering
  • Mathematics / statistics
  • Electronic information / automation
  • AI / data science

Stage

  • Undergraduates (second and third year preferred)
  • Master's and PhD students
  • High-school students with strong programming

What you need

  • Solid mathematics: linear algebra, probability, calculus
  • Python / Linux / scientific computing, or willing to learn
  • Comfortable reading English papers

Commitment

  • Regular research training for the duration of the project
  • At least half a day a week for discussion or study
  • A short report or project summary at the end

Most reading is in English. Working out what to read and which tool to use is part of the training.