About Us

The UCLA Astrophysics Data Lab aims to:

  • Answer fundamental questions in astrophysics using the increasingly large datasets that are available now and in the future.
  • Build the framework for translating machine learning methods to astrophysics.
  • Develop innovative ways of using data for astrophysics.
  • Help astronomers integrate new data-driven methods and practices into their work.

Cosmology with machine learning

Scientifically, the Lab is currently concentrating on probabilistic neural networks as a method for improving photometric redshifts to measure dark energy with the upcoming LSST survey with the Vera Rubin Observatory. The nature of dark energy is one of the biggest mysteries in physics and astrophysics as it represents 72% of the energy density of the Universe but we have very few constraints on its physical origin.

Because dark energy affects the fabric of the Universe, the LSST survey will image over 5 billion galaxies allowing us to see the effects of dark energy on the distribution of matter at unprecedented precision. For this project to be successful, we need to know the distances to all these galaxies and understand the uncertainty of our measurements.

Because galaxies are very complex objects, the most promising methods involve using machine learning to empirically learn about how galaxies vary across cosmic time. The biggest challenge for these machine learning models is to have accurate uncertainty estimates. Our novel approach to this problem is to use probabilistic neural networks, which are like more typical neural networks but can propagate uncertainties throughout the model.

Probabilistic deep learning such as Bayesian Neural Networks has many advantages compared to traditional neural networks, including: better uncertainty representations, better point predictions, and offers better interpretability of neural networks because they can be viewed through the lens of probability theory. In this way we can draw upon decades of development in Bayesian inference development. We think that probabilistic neural networks have the potential to be a transformative tool for cosmology by embedding both physics of the universe and telescope and instrumental characteristics into the same neural network.

The human side of machine learning

An important aspect of the Astrophysics Data Lab is to adapt machine learning tools from industry for scientific purposes and to help develop best practices in their use. We are currently investing the following research questions: (1) how do new computational capabilities change practices of doing science? (2) What knowledge and skills are needed to shift to new computational environments? (3) How do training, labor, credit, buy-in and rewards accelerate or hinder transitions to new workflows?

To answer these questions, have been interviewing astronomers using machine learning to understand their motivation for using machine learning and how they are using it in their own work. We have interviewed astronomers from around the world and at different stages in their career including grad students, postdocs, and faculty.

We have also done several projects aiming at replicating machine learning papers to understand the issues involved in machine learning workflows and how to preserve machine learning models. We are currently preparing a paper on this work and aim to disseminate it at machine learning conferences focusing on the physical sciences.

  • Prof. Tuan Do (UCLA)
    Assistant Professor, Astronomy
    Office: PAB 3-945, Phone: 310-794-9466
    Email: tdo@astro.ucla.edu

  • Prof. Bernadette (Bernie) Boscoe
    Visiting Assistant Professor at Occidental College, previously postdoctoral fellow at UCLA

  • Evan Jones
    UCLA graduate student

  • Christine (Chris) Ma
    UCLA undergraduate student
  • Kevin Alfaro
    UCLA undergraduate student
  • Nissia Indradjaja
    UCLA undergraduate student
  • Yunqi (Billy) Li
    UCLA undergraduate student
  • Zooey Nguyen
    UCLA undergraduate student

Previous Group Members
  • Yujie Wan - undergraduate student

Our work is partially supported by the Alfred P. Sloan Foundation. We thank our funders for their generous support.