A5100339674- Gamma-ray bursts and supernovae
- Statistical and numerical algorithms
- Galaxies: Formation, Evolution, Phenomena
- Astronomy and Astrophysical Research
- Gaussian Processes and Bayesian Inference
- Multidisciplinary Science and Engineering Research
- Stellar, planetary, and galactic studies
- Pulsars and Gravitational Waves Research
- Geophysics and Gravity Measurements
- Big Data Technologies and Applications
- Cell Image Analysis Techniques
- Parallel Computing and Optimization Techniques
- Radio Astronomy Observations and Technology
- Advanced Data Storage Technologies
- Astronomical Observations and Instrumentation
- Time Series Analysis and Forecasting
- Astrophysical Phenomena and Observations
- Adaptive optics and wavefront sensing
- Cosmology and Gravitation Theories
- Viral Infectious Diseases and Gene Expression in Insects
- High-Energy Particle Collisions Research
- Advanced Scientific Research Methods
- Scientific Research and Discoveries
- Monoclonal and Polyclonal Antibodies Research
- Neural Networks and Applications
SLAC National Accelerator Laboratory
2021-2024
Kavli Institute for Particle Astrophysics and Cosmology
2021-2023
Stanford University
2020-2023
Therapeutic antibody design is a complex multi-property optimization problem that traditionally relies on expensive search through sequence space. Here, we introduce "Lab-in-the-loop," paradigm shift for orchestrates generative machine learning models, multi-task property predictors, active ranking and selection, in vitro experimentation semi-autonomous, iterative loop. By automating the of variants, prediction, selection designs to assay lab, ingestion data, enable holistic, end-to-end...
Abstract We describe the simulated sky survey underlying second data challenge (DC2) carried out in preparation for analysis of Vera C. Rubin Observatory Legacy Survey Space and Time (LSST) by LSST Dark Energy Science Collaboration (LSST DESC). Significant connections across multiple science domains will be a hallmark LSST; DC2 program represents unique modeling effort that stresses this interconnectivity way has not been attempted before. This encompasses full end-to-end approach: starting...
ABSTRACT In recent years, breakthroughs in methods and data have enabled gravitational time delays to emerge as a very powerful tool measure the Hubble constant H0. However, published state-of-the-art analyses require of order 1 yr expert investigator up million hours computing per system. Furthermore, precision improves, it is crucial identify mitigate systematic uncertainties. With this delay lens modelling challenge, we aim assess level accuracy techniques that are currently fast enough...
We investigate the use of approximate Bayesian neural networks (BNNs) in modeling hundreds time-delay gravitational lenses for Hubble constant ($H_0$) determination. Our BNN was trained on synthetic HST-quality images strongly lensed active galactic nuclei (AGN) with lens galaxy light included. The can accurately characterize posterior PDFs model parameters governing elliptical power-law mass profile an external shear field. then propagate BNN-inferred into ensemble $H_0$ inference, using...
Abstract Quasars are bright and unobscured active galactic nuclei (AGN) thought to be powered by the accretion of matter around supermassive black holes at centers galaxies. The temporal variability a quasar’s brightness contains valuable information about its physical properties. UV/optical is stochastic process, often represented as damped random walk described differential equation (SDE). Upcoming wide-field telescopes such Rubin Observatory Legacy Survey Space Time (LSST) expected...
Algorithms for machine learning-guided design, or design algorithms, use learning-based predictions to propose novel objects with desired property values. Given a new task -- example, proteins high binding affinity therapeutic target one must choose algorithm and specify any hyperparameters predictive and/or generative models involved. How can these decisions be made such that the resulting designs are successful? This paper proposes method selection, which aims select algorithms will...
Abstract Large-scale astronomical surveys have the potential to capture data on large numbers of strongly gravitationally lensed supernovae (LSNe). To facilitate timely analysis and spectroscopic follow-up before supernova fades, an LSN needs be identified soon after it begins. quickly identify LSNe in optical survey sets, we designed ZipperNet, a multibranch deep neural network that combines convolutional layers (traditionally used for images) with long short-term memory time series). We...
Abstract In the past few years, approximate Bayesian Neural Networks (BNNs) have demonstrated ability to produce statistically consistent posteriors on a wide range of inference problems at unprecedented speed and scale. However, any disconnect between training sets distribution real-world objects can introduce bias when BNNs are applied data. This is common challenge in astrophysics cosmology, where unknown our universe often science goal. this work, we incorporate with flexible posterior...
Abstract We present a Bayesian graph neural network (BGNN) that can estimate the weak lensing convergence ( κ ) from photometric measurements of galaxies along given line sight (LOS). The method is particular interest in strong gravitational time-delay cosmography (TDC), where characterizing “external convergence” ext lens environment and LOS necessary for precise Hubble constant H 0 inference. Starting large-scale simulation with resolution ∼1′, we introduce fluctuations on galaxy–galaxy...
lenstronomy is an Astropy-affiliated Python package for gravitational lensing simulations and analyses. was introduced by Birrer Amara (2018) based on the linear basis set approach et a. (2015). The user developer base of has substantially grown since then, software become integral part a wide range recent analyses, such as measuring Hubble constant with time-delay strong or constraining nature dark matter from resolved unresolved small scale distortion statistics. modular design allowed...
In preparation for cosmological analyses of the Vera C. Rubin Observatory Legacy Survey Space and Time (LSST), LSST Dark Energy Science Collaboration (LSST DESC) has created a 300 deg$^2$ simulated survey as part an effort called Data Challenge 2 (DC2). The DC2 sky survey, in six optical bands with observations following reference observing cadence, was processed Pipelines (19.0.0). this Note, we describe public data release resulting object catalogs coadded images five years along...
Quasars are bright and unobscured active galactic nuclei (AGN) thought to be powered by the accretion of matter around supermassive black holes at centers galaxies. The temporal variability a quasar's brightness contains valuable information about its physical properties. UV/optical is stochastic process, often represented as damped random walk described differential equation (SDE). Upcoming wide-field telescopes such Rubin Observatory Legacy Survey Space Time (LSST) expected observe tens...
Quasars are bright active galactic nuclei powered by the accretion of matter around supermassive black holes at center galaxies. Their stochastic brightness variability depends on physical properties disk and hole. The upcoming Rubin Observatory Legacy Survey Space Time (LSST) is expected to observe tens millions quasars, so there a need for efficient techniques like machine learning that can handle large volume data. Quasar believed be driven an X-ray corona, which reprocessed emitted as...
Among the most extreme objects in Universe, active galactic nuclei (AGN) are luminous centers of galaxies where a black hole feeds on surrounding matter. The variability patterns light emitted by an AGN contain information about physical properties underlying hole. Upcoming telescopes will observe over 100 million multiple broadband wavelengths, yielding large sample multivariate time series with long gaps and irregular sampling. We present method that reconstructs simultaneously infers...
We present a Bayesian graph neural network (BGNN) that can estimate the weak lensing convergence ($\kappa$) from photometric measurements of galaxies along given line sight. The method is particular interest in strong gravitational time delay cosmography (TDC), where characterizing "external convergence" ($\kappa_{\rm ext}$) lens environment and sight necessary for precise inference Hubble constant ($H_0$). Starting large-scale simulation with $\kappa$ resolution $\sim$1$'$, we introduce...