Explore simulation-based inference techniques for gravitational wave astronomy in this comprehensive lecture by Kyle Cranmer at IPAM's Workshop III. Delve into the taxonomy of recent developments, including frequentist vs. Bayesian approaches, learned and engineered summary statistics, densities vs. density ratios, and amortized vs. sequential methods. Gain insights into how these techniques apply to gravitational wave astronomy and the role of inductive bias in neural network-based approaches. Examine issues in inference, high-fidelity simulators, simulator shortcomings, and examples from particle physics. Investigate Approximate Bayesian Computation, dimensionality challenges, and the frontier of simulation-based inference. Learn about neural networks, deep learning, binary classifiers, unsupervised learning techniques, and their applications in population-level inference for gravitational wave astronomy.
Simulation-Based Inference for Gravitational Wave Astronomy - IPAM at UCLA
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Syllabus
Introduction
Issues in Inference
High fidelity simulators
Simulator shortcomings
Simulator examples
Particle physics example
Approximate Bayesian Computation
Dimensionality
frontier of simulationbased inference
Simulationbased inference taxonomy
Simulationbased inference workflow
Neural networks
Deep learning
Binary classifiers
Workflow
Unsupervised Learning
Techniques
Training Data
Population Level Inference
Other Examples
Taught by
Institute for Pure & Applied Mathematics (IPAM)
